SAS 9.

4®

Graph Template Language

Reference
Fourth Edition

SAS® Documentation
The correct bibliographic citation for this manual is as follows: SAS Institute Inc. 2015. SAS® 9.4 Graph Template Language: Reference, Fourth
Edition. Cary, NC: SAS Institute Inc.
SAS® 9.4 Graph Template Language: Reference, Fourth Edition
Copyright © 2015, SAS Institute Inc., Cary, NC, USA

All rights reserved. Produced in the United States of America.

For a hard-copy book: No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, or otherwise, without the prior written permission of the publisher, SAS Institute Inc.
For a web download or e-book: Your use of this publication shall be governed by the terms established by the vendor at the time you acquire this
publication.
The scanning, uploading, and distribution of this book via the Internet or any other means without the permission of the publisher is illegal and
punishable by law. Please purchase only authorized electronic editions and do not participate in or encourage electronic piracy of copyrighted
materials. Your support of others' rights is appreciated.
U.S. Government License Rights; Restricted Rights: The Software and its documentation is commercial computer software developed at private
expense and is provided with RESTRICTED RIGHTS to the United States Government. Use, duplication or disclosure of the Software by the
United States Government is subject to the license terms of this Agreement pursuant to, as applicable, FAR 12.212, DFAR 227.7202-1(a), DFAR
227.7202-3(a) and DFAR 227.7202-4 and, to the extent required under U.S. federal law, the minimum restricted rights as set out in FAR 52.227-19
(DEC 2007). If FAR 52.227-19 is applicable, this provision serves as notice under clause (c) thereof and no other notice is required to be affixed to
the Software or documentation. The Government's rights in Software and documentation shall be only those set forth in this Agreement.
SAS Institute Inc., SAS Campus Drive, Cary, North Carolina 27513-2414.
July 2015
SAS® and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other
countries. ® indicates USA registration.
Other brand and product names are trademarks of their respective companies.
Contents

What’s New in SAS 9.4 Graph Template Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

PART 1 Fundamentals 1

Chapter 1 • Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Graph Template Language (GTL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Basic Anatomy of an ODS Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Graphical Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Legends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Flexible Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
About the Examples in This Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Examples and Resources on the Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

PART 2 Graph Block 19

Chapter 2 • BEGINGRAPH Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

PART 3 Layout Statements 39

Chapter 3 • Summary of Layout Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Single-cell Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Multi-cell Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Data-driven Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Legend Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Chapter 4 • Layout Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

PART 4 Plot Statements 173

Chapter 5 • Key Concepts for Using Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

Minimum Requirements to Generate a Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
ODS Graphics Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Display Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Location and Position of Curve Labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
iv Contents

Chapter 6 • Plot Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190

PART 5 Plot Axes 873

Chapter 7 • Axis Features in Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875
How Axis Features Are Determined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875
Controlling Axis Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883

Chapter 8 • Axis Options in Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 889

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 889

PART 6 Legend Statements 1087

Chapter 9 • Legend Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1089

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1089

PART 7 Text Statements 1135

Chapter 10 • Managing Text Items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1137

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1137
Using Prefix Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1138
Using Text Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1140
Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1142

Chapter 11 • Text Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1147

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1147

PART 8 Custom Marker Definition Statements 1171

Chapter 12 • Custom Marker Definition Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1173

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1173

PART 9 Draw Statements 1187

Chapter 13 • Key Concepts for Using Draw Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1189

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1189
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1189
Types of Elements That Can Be Drawn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1192
About the Drawing Space and Drawing Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1192
How the Drawn Elements Are Anchored . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195
About Drawing Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195
Contents v

Chapter 14 • Draw Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1199

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1199

PART 10 GTL Annotation Facility 1259

Chapter 15 • About the GTL Annotation Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1261

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1261
The Annotation Data Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1262
The ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1266
The SGRENDER Statement SGANNO Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1266

Chapter 16 • The ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1267

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1267

PART 11 Attribute Maps 1275

Chapter 17 • Key Concepts for Using Attribute Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1277

About Attribute Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1277
Defining a Discrete Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1278
Defining a Range Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1281
Referencing an Attribute Map in Your Plot Statements . . . . . . . . . . . . . . . . . . . . . . . 1282

Chapter 18 • Discrete Attribute Map Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1287

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1287

Chapter 19 • Range Attribute Map Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1301

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1301

PART 12 Run-Time Programming Features 1311

Chapter 20 • Dynamics and Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1313

Template Types on PROC TEMPLATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1313
DYNAMIC, MVAR, and NMVAR Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1313
Dynamics Compared to Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1315

Chapter 21 • Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1317

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1317
GTL Expressions Compared to WHERE Expressions . . . . . . . . . . . . . . . . . . . . . . . . 1318
An Expression in Statement Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1318

Chapter 22 • Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1321

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1321
SAS Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1321
Functions Defined Only in GTL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1324
GTL Summary Statistic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1328

Chapter 23 • Conditional Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1333

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1333
Conditional Logic Determines Statement Rendering . . . . . . . . . . . . . . . . . . . . . . . . . 1334
vi Contents

GTL Does Not Provide ELSE IF Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1335

PART 13 Appendixes 1337

Appendix 1 • Syntax Conventions and Argument Value Types . . . . . . . . . . . . . . . . . . . . . . . . 1339

Syntax Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1339
Value Types for Statement Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1339

Appendix 2 • Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1343

Appendix 3 • Display Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347

General Syntax for Attribute Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347
Attributes Available for the Attribute Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1348
Available Line Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1352

Appendix 4 • SAS Formats Not Supported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353

Using SAS Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353
Unsupported Numeric Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353
Unsupported Date and Time Formats Related to ISO 8601 . . . . . . . . . . . . . . . . . . . . 1354
Other Unsupported Date and Time Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1354
Unsupported Currency Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1355
Unsupported User-Defined Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1355

Appendix 5 • Generalized Macro for BOXPLOTPARM Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 1357

Appendix 6 • Memory Management for ODS Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363

SAS Options Affecting Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363
Managing a Java Out of Memory Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363

Recommended Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1367

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1369
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1381
vii

What’s New in SAS 9.4 Graph

Template Language

Overview

New and enhanced statements for the Graph Template Language (GTL) extend the
versatility of the language and introduce new plot types. The following changes are
included in this release:
• new plot statements
• new features for general use
• enhancements to SAS 9.3 statements

Possible Changes Required for Your Existing

SAS GTL Programs

There are changes in GTL for SAS 9.4 and subsequent maintenance releases that might
require you to modify your existing SAS GTL programs. These changes can impact SAS
GTL programs that use the DENSITYPLOT or BARCHART statements. Review this
section carefully to determine whether you need to modify your existing SAS GTL
programs for SAS 9.4.

Programs That Use the BARCHART Statement

In SAS 9.4, the BARCHART statement STAT=PCT option now displays percentages in
the range 0–100 in order to be consistent with other GTL statements. In prior releases,
STAT=PCT displays proportional values in the range 0–1. To restore the proportional
values in SAS 9.4 and later releases, change STAT=PCT to STAT=PROPORTION.
In the third maintenance release of SAS 9.4, the COLORSTAT= option is added to the
BARCHART statement. It is enabled by the COLORRESPONSE= option. The
COLORSTAT= option specifies the statistic to be calculated for the data range of the
bar-color gradient. The default is SUM. For existing SAS programs that use the
BARCHART statement, if STAT= is used with COLORRESPONSE= in the
BARCHART statement and STAT= specifies a statistic other than SUM, then the bar-
chart colors and color statistic might change from those of previous releases. In that
case, to restore the original colors and color statistic, set COLORSTAT= in the
BARCHART statement to the same statistic that is specified in STAT=.
See BARCHART in “Plot Statement Enhancements” on page xiv.
viii Graph Template Language

Programs That Use the DENSITYPLOT Statement

In SAS 9.4, the WEIGHT= option of the KERNEL( ) distribution option in the
DENSITYPLOT statement is changed to WEIGHTFUNCTION=. This change enables
the addition of the WEIGHT= option in the DENSITYPLOT statement. Starting with
SAS 9.4, the WEIGHT= option is not valid in the KERNEL() distribution option and
results in a syntax error. If your existing SAS programs specify the WEIGHT= option in
the KERNEL( ) distribution option, then you must change it to WEIGHTFUNCTION=.
See DENSITYPLOT in “Plot Statement Enhancements” on page xiv.

Programs That Use the HEATMAPPARM Statement

In the third maintenance release of SAS 9.4, the DISCRETEX= and DISCRETEY=
options are added to the HEATMAPPARM statement. These options specify whether the
X or Y axis is discrete. The default is FALSE. For existing SAS programs that use the
HEATMAPPARM statement, if the X axis is discrete, the DISCRETEX=TRUE option
must be specified in the HEATMAPPARM statement. Likewise, if the Y axis is discrete,
the DISCRETEY=TRUE option must be specified in the HEATMAPPARM statement.
Otherwise, the heat map might not be drawn.

Programs That Use the WATERFALLCHART Statement

In the third maintenance release of SAS 9.4, the COLORSTAT= option is added to the
WATERFALLCHART statement. It is enabled by the COLORRESPONSE= option. The
COLORSTAT= option specifies the statistic to be calculated for the data range of the
color gradient. The default is SUM. For existing SAS programs that use the
WATERFALLCHART statement, if STAT= is used with COLORRESPONSE= in the
WATERFALLCHART statement and STAT= specifies a statistic other than SUM, then
the chart colors and color statistic might change from those of previous releases. In that
case, to restore the original colors and color statistic, set COLORSTAT= in the
WATERFALLCHART statement to the same statistic that is specified in STAT=. See
WATERFALLCHART in “Plot Statement Enhancements” on page xiv.

New Plot Statements

The following plot statements are new:

• ANNOTATE draws annotations from annotation instructions that are stored in a SAS
data set. It can draw all of the annotations in the data set or a subset only.
• AXISTABLE draws textual values (character or numeric) on the graph at locations
that are aligned with the X or Y axis.
• LINECHART creates a summarized line chart that is computed from input data.
In the first maintenance release of SAS 9.4, the following statement is new:
• POLYGONPLOT draws one or more polygons from data that is stored in a SAS data
set.
In the second maintenance release of SAS 9.4, the following statement is new:
Color-Priority Graph Data Attribute Rotation ix

• TEXTPLOT displays text values at specific X and Y locations in the graph. For
information about subsequent enhancements to this statement, see “Plot Statement
Enhancements” on page xiv.
In the third maintenance release of SAS 9.4, the following statement is new:
• HEATMAP creates a plot of color-coded rectangles for the response variable of a
pair of X and Y variables after it bins the data in two dimensions.

New Legend Statement

In the third maintenance release of SAS 9.4, the LEGENDTEXTITEMS statement is

new. This statement creates the definition for data-driven text items that can be included
in a discrete legend. The items and optional labels are stored in the plot data set.

New Marker Definition Statements

In the first maintenance release of SAS 9.4, the following statements are new:
• SYMBOLCHAR defines a marker symbol from a Unicode character value.
• SYMBOLIMAGE defines a marker symbol from a GIF, JPG, or PNG image that is
stored in a file.
You can use these statements to define custom marker symbols for your graphs. For
information about subsequent enhancements to these statements, see “Marker Definition
Statement Enhancements” on page xxxvii.

New Function

In the first maintenance release of SAS 9.4, the TYPEOF function is new. This function
returns the type (numeric or character) of a specified column at run time. You can use the
TYPEOF function to take specific actions in your template at run time based on the input
data type.

New Features for General Use

Color-Priority Graph Data Attribute Rotation

The GTL now supports a color-priority rotation pattern for cycling graph data attributes.
In the default rotation pattern, graph data attributes rotate as they are defined in the
GraphData1–GraphDataN style elements. The new color-priority rotation pattern first
cycles through all of the colors while holding the marker symbol, line pattern, or fill
x Graph Template Language

pattern constant. When all of the colors have been used, it increments to the next marker
symbol, line pattern, or fill pattern, and then repeats the colors.
Color-priority rotation is enabled in one of the following ways:
• the currently active style sets the Graph:attrPriority attribute to COLOR
• the ATTRPRIORITY=COLOR option is specified in the ODS GRAPHICS
statement, which overrides the Graph:attrPriority style attribute
• the ATTRPRIORITY=COLOR option is specified in the BEGINGRAPH statement
of the template, which overrides the ODS GRAPHICS ATTRPRIORITY= option for
all plots in that template
When none of these conditions are true, the default rotation pattern is used. For more
information, see “BEGINGRAPH Statement Enhancements” on page xi.

Template-Level Graph Data Attribute Overrides

The GTL provides new options that enable you to override GraphData1–GraphDataN
style attributes for all of the plots within a template. They provide an easy way to modify
group attributes without having to create a custom ODS style. These options are used in
the BEGINGRAPH statement of a template. By using these options, you can override
attributes in the GraphData1–GraphDataN style elements by specifying one or more of
the following lists for all of the plots within a template:
• a list of colors that replace the graph data colors (fill) in the current style
• a list of contrast colors that replace the graph data contrast colors in the current style
• a list of marker symbols that replace the graph data marker symbols in the current
style
• a list of line patterns that replace the graph data line patterns in the current style
In addition, you can specify a priority for the group value attribute cycling. For more
information, see “BEGINGRAPH Statement Enhancements” on page xi.

Outer Padding
The GTL provides a new OUTERPAD= option that enables you to control the padding
around the outside of layouts, legends, titles, footnotes, and text entries. You can use this
option to modify the outer padding when the default padding is not sufficient. You can
specify a single padding value for all four sides, or you can specify different values for
each side.

Unicode Values in User-Defined Formats

Starting with the third maintenance release of SAS 9.4, ODS Graphics supports Unicode
values in user-defined formats. The Unicode value must be escaped with the (*ESC*)
escape sequence, as shown in the following example.
"(*ESC*){unicode beta}"

ODS Graphics does not support the use of a user-defined ODS escape character to
escape Unicode values in user-defined formats.
BEGINGRAPH Statement Enhancements xi

General Enhancements Supported by Many of the Plots

The following new features that are supported by many of the plot statements are worth
highlighting. The individual plot statements that support these features are identified in
“Plot Statement Enhancements” on page xiv.
• The algorithm that is used to place data labels has been improved to more effectively
position the data labels in the vicinity of their data markers while avoiding label
collisions.
• In certain cases, you can rotate or split data labels, curve labels, and discrete-axis tick
values in order to fit them in the available space.
• Subpixel rendering is now supported, which produces smoother curves and more
precise bar spacing in plots.
• You can now suppress tips and URLs from individual plots.

Enhancements to SAS 9.4 Statements

BEGINGRAPH Statement Enhancements

• ATTRPRIORITY= specifies a priority for cycling of the attributes for group
attributes.
• DATACOLORS= specifies the list of fill colors that replaces the graph data colors
from the GraphData1–GraphDataN style elements.
• DATACONTRASTCOLORS= specifies the list of contrast colors that replaces the
graph data contrast colors from the GraphData1–GraphDataN style elements.
• DATALINEPATTERNS= specifies the list of line patterns that replaces the graph
data line patterns from the GraphData1–GraphDataN style elements.
• DATASKIN= enhances the visual appearance of all plots in the template that support
data skins.
• DATASYMBOLS= specifies the list of marker symbols that replaces the graph data
marker symbols from the marker symbols that are defined in the GraphData1–
GraphDataN style elements.
• SUBPIXEL= specifies whether subpixel rendering is used for drawing smooth
curved lines or for spacing bars more precisely.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• AXISBREAKSYMBOL= specifies a symbol to use on the axis lines to indicate a
break in the axis.
• AXISBREAKTYPE= specifies whether the axis break is indicated in the full
display or only on the axis line.
• AXISLINEEXTENT= specifies the extent of the axis line for all axes.
xii Graph Template Language

• DISCRETEAXISOFFSETPAD= specifies whether additional padding is added

to the minimum and maximum axis offsets for discrete axes.
• OPAQUE= specifies whether the graph background is opaque or transparent.

Layout Statement Enhancements

LAYOUT DATALATTICE and LAYOUT DATAPANEL:
• For cell insets, match-merging is now supported for merging the inset and analysis
data. The DATASCHEME= suboption of the INSETOPTS= option specifies whether
one-to-one merging or match merging was used to merge the data.
• HEADERLABELDISPLAY= now supports NONE, which suppresses the row and
column headings, or the cell headings.
• SORTORDER= specifies the order of the cells along the columns and rows.
• LAYOUT DATALATTICE only:
• COLUMNDATARANGE=, ROWDATARANGE=, COLUMN2DATARANGE=,
and ROW2DATARANGE= now support AUTO, which selects the range
automatically, based on the column weight or row weight, and the axis type.
• COLUMNWEIGHT= specifies how to assign weights to the columns widths.
• ROWWEIGHT= specifies how to assign weights to the row heights.
• Starting with the first maintenance release of SAS 9.4, the following enhancements
are valid:
• HEADERPACK= specifies whether the values listed in the cell headers are
displayed as a delimited list in a single header cell in order to save space.
• HEADERSEPARATOR= specifies a separator to place between each value in the
cell header.
• HEADERSPLITCOUNT= specifies the number of class variables in the cell
header after which the cell header wraps to a separate line.
• Starting with the second maintenance release of SAS 9.4, the following
enhancements are valid:
• For the INSETOPTS= option:
• CONTENTDISPLAY= specifies whether the variable information displayed
in the inset includes the column label and value, or the column value only.
• SEPARATOR= specifies a new separator for the column label and value.
• Starting with the third maintenance release of SAS 9.4, HEADERBORDER=
specifies whether a border is drawn around the header cells.
LAYOUT GLOBALLEGEND:
• Starting with the first maintenance release of SAS 9.4, the default padding between
the global legend and the plot area (including the axes) is 10 pixels. The
OUTERPAD= option can be used to adjust the padding if necessary.
LAYOUT OVERLAY:
• INNERMARGIN statement:
• ALIGN= now supports LEFT and RIGHT in addition to TOP and BOTTOM.
• BACKGROUNDCOLOR= specifies the color of the inner margin background.
Legend Statement Enhancements xiii

• OPAQUE= specifies whether the inner margin's background is opaque.

• Starting with the first maintenance release of SAS 9.4, the following
enhancements are valid:
• SEPARATOR= specifies whether a separating line is drawn between the
inner margin and the rest of the layout content.
• SEPARATORATTRS= specifies the attributes of the inner margin separating
line.
• Starting with the second maintenance release of SAS 9.4, the following
enhancements are valid:
• GUTTER= specifies the gap between stacked items in the inner margin.
• PAD= specifies the amount of extra space that is added inside the inner
margin border.
LAYOUT OVERLAYEQUATED:
• The HEATMAPPARM statement is now supported.
• EQUATETYPE= now supports SQUAREDATA, which specifies that both the X and
Y axes have the same range, but can have different tick values.
LAYOUT PROTOTYPE:
• The HEATMAPPARM statement is now supported.
• Starting with the first maintenance release of SAS 9.4, the INNERMARGIN
statement is now supported in a LAYOUT PROTOTYPE block.

Legend Statement Enhancements

AXISLEGEND:
• Starting with the second maintenance release of SAS 9.4, the BORDER= option
defaults to style reference GraphLegendBackground:FrameBorder.
DISCRETELEGEND:
• Starting with the first maintenance release of SAS 9.4, the following enhancements
are valid:
• The default padding between the legend and the plot area (including the axes) is
10 pixels, depending on the context. You can use the OUTERPAD= option to
adjust the padding if necessary.
• ITEMSIZE= specifies the size of specific types of items in a discrete legend.
• Starting with the second maintenance release of SAS 9.4, the BORDER= option
defaults to style reference GraphLegendBackground:FrameBorder.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• FILLITEMOUTLINE= specifies whether the fill swatches are outlined only
when enabled by the contributing statements or are always outlined.
• The ITEMSIZE= option now supports the following suboptions:
• FILLASPECTRATIO= specifies the aspect ratio for the fill swatches.
• FILLHEIGHT= specifies the height of the fill swatches.
xiv Graph Template Language

• HEIGHTSCALE= specifies a scaling factor that is to be applied to the fill

swatch height.
• SORTBY= specifies whether text legend items are sorted by label or by text.
CONTINUOUSLEGEND:
• EXTRACTSCALE= specifies whether to extract a scale factor from the tick values
and use it to reduce the tick value width.
• EXTRACTSCALETYPE= specifies whether to extract a named scale or a scientific-
notation scale.
• Starting with the second maintenance release of SAS 9.4, the BORDER= option
defaults to style reference GraphLegendBackground:FrameBorder.
• Starting with the third maintenance release of SAS 9.4, INTEGER= specifies
whether only integer tick values are used in the continuous legend.
LEGENDITEM:
• FILLEDOUTLINEDMARKERS= specifies whether markers are drawn with both
fills and outlines.
• Starting with the third maintenance release of SAS 9.4, FILLDISPLAY= specifies
whether the fill swatch for this legend item displays fill only or displays fill and
outline.
MERGEDLEGEND:
• Starting with the first maintenance release of SAS 9.4, the following enhancements
are valid:
• ADDITIONALNAMES= specifies the name of one or more sources for legend
items that are to be added to the merged legend after merging takes place.
• ITEMSIZE= specifies the size of specific types of items in a merged legend.
• Starting with the second maintenance release of SAS 9.4, the BORDER= option
defaults to style reference GraphLegendBackground:FrameBorder.

Plot Statement Enhancements

AXISTABLE:
• Starting with the second maintenance release of SAS 9.4, the following
enhancements are valid:
• CLASSDISPLAY= specifies how the class values are displayed.
• CLASSORDER= specifies the order in which the class values are displayed.
• CLUSTERWIDTH= specifies the width of the group clusters as a fraction of the
midpoint spacing.
• DISPLAY= now supports VALUES, which displays the table values.
• DROPONMISSING= specifies whether the entire axis table is dropped when all
of the VALUE= column values are missing.
• GUTTER= specifies the gap between rows when a class variable is used.
• INCLUDEMISSINGCLASS= specifies whether missing class values are
represented in the table.
• LABEL= specifies the text for the table label.
Plot Statement Enhancements xv

• LABELHALIGN= specifies the horizontal alignment of the column labels

relative to the column width in a Y-axis table.
• LABELJUSTIFY= specifies the justification of the column label, when
displayed.
• PAD= specifies the amount of extra space that is added inside the table border.
• SHOWMISSING= specifies whether missing values are represented in the table.
• TITLE= specifies the text for the table title.
• TITLEATTRS=specifies the color and font attributes of the table title.
• TITLEHALIGN= specifies the horizontal alignment of the axis table header label
relative to the table width for a Y-axis table.
• TITLEJUSTIFY= specifies the justification of the title string.
• VALUEFORMAT= specifies a SAS format or a user-defined format for the table
values.
• VALUEHALIGN= specifies the horizontal alignment of the column values
relative to the column width in a Y-axis table.
• VALUEJUSTIFY= specifies the justification of the values in the axis table.
• The following options are replaced and considered deprecated:
• HEADERLABEL= is replaced with TITLE=.
• HEADERLABELATTRS= is replaced with TITLEATTRS=.
The deprecated options are still honored, but the new options are the preferred
options.
• Starting with the third maintenance release of SAS 9.4, TITLEHALIGN= specifies
the horizontal alignment of the axis table header label relative to the table width for
Y-axis tables and X-axis tables.
BANDPLOT:
• TIP= now supports NONE, which suppressed data tips from the plot.
• Starting with the second maintenance release of SAS 9.4:
• ANTIALIAS= specifies whether anti-aliasing is turned off for a band plot.
BARCHART:
• X= is changed to CATEGORY= in order to be consistent with the other GTL plot
statements. The X= option is still valid for backward compatibility. However, you
should change the X= option to CATEGORY= in your BARCHART statements.
Note: If you specify X as a data tip role when you change X= to CATEGORY=, then
you must also change X to CATEGORY in your data tip options.
• Y= is changed to RESPONSE= in order to be consistent with the other GTL plot
statements. The Y= option is still valid for backward compatibility. However, you
should change the Y= option to RESPONSE= in your BARCHART statements.
Note: If you specify Y as a data tip role when you change Y= to RESPONSE=, then
you must also change Y to RESPONSE in your data tip options.
• BARLABELFITPOLICY= specifies a policy for avoiding collisions among the bar
labels when bar labels are displayed.
• BASELINEATTRS= specifies the appearance of the baseline.
xvi Graph Template Language

• The baseline is now drawn by default. To suppress the baseline, use the
BASELINEATTRS= option to set the line thickness to 0.
• DATALABELFITPOLICY= specifies a policy for avoiding collisions among the bar
labels when bar labels are displayed.
• STAT= now supports PROPORTION, which displays proportions in the range 0–1.
• STAT=PCT now displays percentages in the range 0–100 in order to be consistent
with other GTL statements.
Note: In prior SAS releases, the BARCHART statement STAT=PCT option displays
proportional values in the range 0–1. To restore the original STAT=PCT results in
SAS 9.4 and later releases, specify STAT=PROPORTION in your BARCHART
statements instead.
• TIP= now supports NONE, which suppressed data tips and URLs from the plot.
• Starting with the second maintenance release of SAS 9.4, the following
enhancements are valid:
• GROUPORDER= supports REVERSEDATA, which orders the groups within a
category in the reverse group-column data order.
• FILLTYPE= specifies whether the fill color is solid or is a gradient that
transitions from fully opaque to fully transparent.
• SEGMENTLABEL= specifies whether a label is displayed inside each bar
segment.
• SEGMENTLABELATTRS= specifies the text properties of the bar segment label
text.
• SEGMENTLABELFITPOLICY= specifies a policy for fitting the bar segment
labels within the bar segments.
• SEGMENTLABELFORMAT= specifies the text format used to display the bar
segment labels.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• The BARCHART statement now supports a linear category axis or a time
category axis.
• COLORBYFREQ= specifies whether the bar colors are based on the frequency
of the category variable.
• COLORMODEL= specifies a color ramp that is to be used with the
COLORRESPONSE= option or the COLORBYFREQ= option.
• COLORRESPONSE= specifies the column or range attribute variable that is
used to map the bar colors.
• COLORSTAT= specifies the statistic to use for computing the response colors.
• DISPLAYZEROLENGTHBAR= specifies whether zero-length bars are drawn.
• GROUP100= displays the computed response values (FREQ, SUM, or MEAN),
normalized to 100%.
• INTERVALBARWIDTH= specifies the width of the bars in an interval bar chart
as a ratio of the interval width.
BARCHARTPARM:
Plot Statement Enhancements xvii

• X= is changed to CATEGORY= in order to be consistent with the other GTL plot

statements. The X= option is still valid for backward compatibility. However, you
should change the X= option to CATEGORY= in your BARCHARTPARM
statements.
• Y= is changed to RESPONSE= in order to be consistent with the other GTL plot
statements. The Y= option is still valid for backward compatibility. However, you
should change the Y= option to RESPONSE= in your BARCHARTPARM
statements.
• BASELINEATTRS= specifies the appearance of the baseline.
• The baseline is now drawn by default. To suppress the baseline, use the
BASELINEATTRS= option to set the line thickness to 0.
• DATALABELFITPOLICY= specifies a policy for avoiding collisions among the bar
labels when bar labels are displayed.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• ERRORBARCAPSHAPE= specifies whether the error bars have a serif cap. Starting
with the second maintenance release of SAS 9.4, this option defaults to style
reference GraphError:CapStyle.
• TIP= now supports NONE, which suppressed data tips and URLs from the plot.
• Starting with the second maintenance release of SAS 9.4, the following
enhancements are valid:
• DATALABELTYPE= specifies whether the data labels display the RESPONSE
values or the values of the column specified by the DATALABEL= option.
• FILLTYPE= specifies whether the fill color is solid or is a gradient that
transitions from fully opaque to fully transparent.
• GROUPORDER= supports REVERSEDATA, which orders the groups within a
category in the reverse group-column data order.
• SEGMENTLABELATTRS= specifies the text properties of the bar segment label
text.
• SEGMENTLABELFITPOLICY= specifies a policy for fitting the bar segment
labels within the bar segments.
• SEGMENTLABELFORMAT= specifies the text format used to display the bar
segment labels.
• SEGMENTLABELTYPE= specifies whether a label is displayed inside each bar
segment.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• The BARCHARTPARM statement now supports a linear category axis or a time
category axis.
• COLORMODEL= specifies a color ramp that is to be used with the
COLORRESPONSE= option.
• COLORRESPONSE= specifies the column or range attribute variable that is
used to map the bar colors.
xviii Graph Template Language

• DISPLAYZEROLENGTHBAR= specifies whether zero-length bars are drawn.

• GROUP100= displays the computed values normalized to 100%.
• INTERVALBARWIDTH= specifies the width of the bars as a ratio of the interval
width.
BLOCKPLOT:
• BLOCKLABEL= specifies alternative text to display for the internal block text
values.
• VALUEFITPOLICY= now supports the SPLIT, SPLITALWAYS, and NONE
policies.
• VALUESPLITCHAR= specifies one or more characters on which the values can be
split if needed.
• VALUESPLITCHARDROP= specifies whether the split characters are included in
the displayed values.
BOXPLOT:
• CAPSHAPE= now supports NONE, which specifies that no shape is displayed at the
ends of the box whiskers.
• DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
• DATASKIN= enhances the visual appearance of the filled boxes.
• DISPLAYSTATS= specifies the statistics that are displayed for each box.
• TIP= now supports NONE, which suppressed data tips from the plot.
• WEIGHT= specifies a column that contains a statistics calculation a priori weight for
each observation of the input data object.
• Starting with the first maintenance release of SAS 9.4, the following enhancements
are valid:
• WHISKERPERCENTILE= specifies the whisker length in percentile units.
• DISPLAYSTATS= now displays the DATAMAX, DATAMIN, and SUMWGT
statistics.
• LEGENDLABEL= now defaults to the Y= column label or name.
• Starting with the second maintenance release of SAS 9.4, the GROUPORDER=
option supports REVERSEDATA, which orders the groups within a category in the
reverse group-column data order.
• Starting with the third maintenance release of SAS 9.4, when DISPLAY= includes
MEAN, the BOXPLOT statement contributes its mean markers to a discrete legend
when TYPE=MARKER is specified in the DISCRETELEGEND statement.
BOXPLOTPARM:
Plot Statement Enhancements xix

• CAPSHAPE= now supports NONE, which specifies that no shape is displayed at the
ends of the box whiskers.
• DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
• DATASKIN= enhances the visual appearance of the filled boxes.
• TIP= now supports NONE, which suppresses data tips from the plot.
• Starting with the first maintenance release of SAS 9.4, the following enhancements
are valid:
• DISPLAYSTATS= now displays the DATAMAX, DATAMIN, and SUMWGT
statistics.
• LEGENDLABEL= now defaults to the Y= column label or name.
• URL= specifies an HTML page that is displayed when a box or an outlier marker
is selected.
• Starting with the second maintenance release of SAS 9.4, the GROUPORDER=
option supports REVERSEDATA, which orders the groups within a category in the
reverse group-column data order.
• Starting with the third maintenance release of SAS 9.4, when DISPLAY= includes
MEAN, the BOXPLOTPARM statement contributes its mean markers to a discrete
legend when TYPE=MARKER is specified in the DISCRETELEGEND statement.
BUBBLEPLOT:
• DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
• TIP= now supports NONE, which suppresses data tips and URLs from the plot.
• Starting with the third maintenance release of SAS 9.4, DRAWORDER= specifies
whether the bubbles are drawn according to bubble size or according to data order.
CONTOURPLOTPARM:
• LEVELS= specifies a list of contour level values.
DENDROGRAM:
• TIP= now supports NONE, which suppresses data tips and URLs from the plot.
DENSITYPLOT:
xx Graph Template Language

• CURVELABELSPLIT= specifies whether to split the curve label at the specified

split characters.
• CURVELABELSPLITCHAR= specifies one or more characters on which the curve
label can be split if needed.
• CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the curve label text.
• CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the curve label block.
• WEIGHT= specifies a column that contains a density-curve calculation a priori
weight for each observation of the input data object.
• The KERNEL() distribution option WEIGHT= is changed to
WEIGHTFUNCTION=. This change enables the addition of the WEIGHT= option
in the DENSITYPLOT statement.
Note: The WEIGHT= option is not valid in the KERNEL() distribution option in
SAS 9.4. If you used the WEIGHT= option in the KERNEL() distribution option
in prior SAS releases, then you must change it to WEIGHTFUNCTION= in SAS
9.4 and later releases.
• Starting with the second maintenance release of SAS 9.4, the following
enhancements are valid:
• GROUP= creates a separate density curve for each unique group value of the
specified column.
• INCLUDEMISSINGGROUP= specifies whether missing values of the group
variable are included in the plot.
DROPLINE:
• DATASKIN= enhances the appearance of the drop line.
• Starting with the first maintenance release of SAS 9.4, DROPTO= now supports
BOTH, which draws one or more drop lines to both the X and Y axes.
ELLIPSE:
Starting with the second maintenance release of SAS 9.4, the following enhancements
are valid:
• GROUP= creates a separate ellipse for each unique group value of the specified
column.
• INCLUDEMISSINGGROUP= specifies whether missing values of the group
variable are included in the plot.
FRINGEPLOT:
• GROUP= creates a distinct set of lines for each unique group value in the specified
column.
• INCLUDEMISSINGGROUP= specifies whether missing values in the group column
are included in the plot.
• INDEX= specifies indices for mapping line attributes (color and line pattern) to one
of the GraphData1–GraphDataN style elements.
• TIP= now supports NONE, which suppresses data tips from the plot.
HEATMAPPARM:
• URL= specifies an HTML page to display when a rectangle is selected.
Plot Statement Enhancements xxi

• TIP= now supports NONE, which suppresses data tips and URLs from the plot.
• Starting with the second maintenance release of SAS 9.4, the
INCLUDEMISSINGCOLOR= option specifies whether missing values of the color-
group variable or the color-response variable are included in the plot.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• DISCRETEX= specifies whether the X axis is discrete when X= specifies a
numeric column.
• DISCRETEY= specifies whether the Y axis is discrete when Y= specifies a
numeric column.
HIGHLOWPLOT:
• CLIPCAP= specifies whether a special clip cap is displayed to indicate where
clipping occurred.
• CLIPCAPSHAPE= specifies the shape of the arrowhead on the clipped end of a line.
• DATASKIN= enhances the visual appearance of the high-low chart filled bars or
lines.
• ENDCAPDISPLAYPOLICY= specifies the policy for displaying end caps when end
caps are present.
• TIP= now supports NONE, which suppresses data tips from the plot.
• Starting with the second maintenance release of SAS 9.4, the GROUPORDER=
option supports REVERSEDATA, which orders the groups within a category in the
reverse group-column data order.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• COLORMODEL= specifies a color ramp that is used with the
COLORRESPONSE= option.
• COLORRESPONSE= specifies the column or range attribute variable that is
used to map the bar or line colors.
HISTOGRAM:
• DATALABELTYPE= specifies the statistic to display at the end of each bar.
• DATASKIN= enhances the visual appearance of the filled bars.
• WEIGHT= specifies a column that contains a bin-width calculation a priori weight
for each observation of the input data object.
• Starting with the first maintenance release of SAS 9.4, the following enhancements
are valid:
• The number of bins is limited to approximately 10,000. When the limit is
exceeded, SAS automatically adjusts the NBINS= or BINWIDTH= value to set
the number of bins to about 10,000.
• DISPLAY= now supports FILLPATTERN.
• FILLPATTERNATTRS= specifies the appearance of the pattern-filled bar area.
• Starting with the second maintenance release of SAS 9.4, the following
enhancements are valid:
• FILLTYPE= specifies whether the fill color is solid or is a gradient that
transitions from fully opaque to fully transparent.
xxii Graph Template Language

• GROUP= creates a separate bar segment or bar for each unique group value of
the specified column.
• INCLUDEMISSINGGROUP= specifies whether missing values of the group
variable are included in the plot.
• The OUTLINEATTRS= option defaults are now consistent with that of
BARCHART.
HISTOGRAMPARM:
• DATALABELFITPOLICY= specifies a policy for avoiding collisions among the bin
labels when bin labels are displayed.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• DATASKIN= enhances the visual appearance of the filled bars.
• TIP= now supports NONE, which suppresses data tips from the plot.
• Starting with the first maintenance release of SAS 9.4, the following enhancements
are valid:
• DISPLAY= now supports FILLPATTERN.
• FILLPATTERNATTRS= specifies the appearance of the pattern-filled bar area.
• Starting with the second maintenance release of SAS 9.4, the FILLTYPE= option
specifies whether the fill color is solid or is a gradient that transitions from fully
opaque to fully transparent.
LINECHART:
• Starting with the second maintenance release of SAS 9.4, the GROUPORDER=
option supports REVERSEDATA, which orders the groups within a category in the
reverse group-column data order.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• COLORMODEL= specifies a color ramp that is used with the
COLORRESPONSE= option.
• COLORRESPONSE= specifies the column or range attribute variable that is
used to map the line, marker, and fill colors.
LINEPARM:
• CURVELABELSPLIT= specifies whether to split the line label at the specified split
characters.
• CURVELABELSPLITCHAR= specifies one or more characters on which the line
label can be split if needed.
• CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the line label text.
• CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the line label block.
LOESSPLOT:
• CURVELABELSPLIT= specifies whether to split the curve label at the specified
split characters.
Plot Statement Enhancements xxiii

• CURVELABELSPLITCHAR= specifies one or more characters on which the curve

label can be split if needed.
• CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the curve label text.
• CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the curve label block.
MODELBAND:
• Starting with the second maintenance release of SAS 9.4:
• ANTIALIAS= specifies whether anti-aliasing is turned off for a model band plot.
• A confidence band that depicts confidence limits for individual predicted values
(CLI) for a weighted spline plot or regression plot is now displayed as a high-low
chart instead of a band.
NEEDLEPLOT:
• BASELINEATTRS= specifies the appearance of the baseline. This option enables
you to suppress the baseline by setting the line thickness to 0.
• DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
• DATASKIN= enhances the appearance of the needle plot lines.
• MARKERATTRS= now supports transparency.
• TIP= now supports NONE, which suppresses data tips and URLs from the plot.
• Starting with the second maintenance release of SAS 9.4, the GROUPORDER=
option supports REVERSEDATA, which orders the groups within a category in the
reverse group-column data order.
PBSPLINEPLOT:
• CURVELABELSPLIT= specifies whether to split the curve label at the specified
split characters.
• CURVELABELSPLITCHAR= specifies one or more characters on which the curve
label can be split if needed.
• CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the curve label text.
• CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the curve label block.
• Starting with the second maintenance release of SAS 9.4, the DEGREE= PBSPLINE
regression option range is 0–10 instead of 0–174.
PIECHART:
• TIP= now supports NONE, which suppresses data tips and URLs from the plot.
xxiv Graph Template Language

• Starting with the first maintenance release of SAS 9.4, CENTERFIRSTSLICE=

specifies whether the first pie slice is centered on the starting angle or starts on it.
• Starting with the third maintenance release of SAS 9.4, suboption
TRANSPARENCY= in the PIECHART statement FILLATTRS= option sets the
transparency of the other slice unless transparency is specified in the
OTHERSLICEOPTS= option.
POLYGONPLOT:
• Starting with the second maintenance release of SAS 9.4:
• ANTIALIAS= specifies whether anti-aliasing is turned off for a polygon plot.
• BACKLIGHT= specifies a back-light effect for the polygon label text.
REFERENCELINE:
• CURVELABELSPLIT= specifies whether to split the reference line label at the
specified split characters.
• CURVELABELSPLITCHAR= specifies one or more characters on which the
reference line label can be split if needed.
• CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the reference line label text.
• CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the reference line label block.
• DATASKIN= enhances the visual appearance of the reference line.
REGRESSIONPLOT:
• CURVELABELSPLIT= specifies whether to split the regression line label at the
specified split characters.
• CURVELABELSPLITCHAR= specifies one or more characters on which the
regression line label can be split if needed.
• CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the regression line label text.
• CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the regression line label block.
• Starting with the second maintenance release of SAS 9.4, the DEGREE= regression
option range is 1–10 instead of 0–174.
SCATTERPLOT:
• CLUSTERAXIS= specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
• DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
Plot Statement Enhancements xxv

• ERRORBARCAPSHAPE= specifies whether the error bars have a serif cap. Starting
with the second maintenance release of SAS 9.4, this option defaults to style
reference GraphError:CapStyle.
• FILLEDOUTLINEDMARKERS= specifies whether markers are drawn with both
fills and outlines.
• JITTER= specifies whether to jitter data markers.
• JITTEROPTS= specifies options for managing jittering.
• LABELSTRIP= specifies whether leading and trailing blanks are stripped from
marker characters or fixed-position data labels before they are displayed in the plot.
• MARKERATTRS= now supports transparency.
• MARKERCHARACTERPOSITION= specifies the justification of the marker
characters.
• MARKERFILLATTRS= specifies the appearance of the filled markers.
• MARKEROUTLINEATTRS= specifies the appearance of the marker outlines.
• TIP= now supports NONE, which suppresses data tips and URLs from the plot.
• Starting with the first maintenance release of SAS 9.4,
OUTLINEDMARKERCHARACTERS= specifies whether backlighting or a drop
shadow is applied to the characters that are used as marker symbols in order to
enhance their appearance in the graph.
• Starting with the second maintenance release of SAS 9.4:
• CONTRIBUTEOFFSETS= specifies whether the plot's space requirements
contribute to the calculation of the axis offsets.
• The GROUPORDER= option supports REVERSEDATA, which orders the
groups within a category in the reverse group-column data order.
• The following options are replaced and considered deprecated:
• MARKERCOLORGRADIENT= is replaced with COLORRESPONSE=.
• MARKERSIZERESPONSE= is replaced with SIZERESPONSE=.
• MARKERSIZEMAX= is replaced with SIZEMAX=.
• MARKERSIZEMIN= is replaced with SIZEMIN=.
The new options are functionally the same as the deprecated options and are
more consistent with the other plot statements. The deprecated options are still
honored, but the new options are the preferred options.
• The OUTLINEDMARKERCHARACTERS= option is deprecated. It is still
honored, but the TEXTPLOT statement is now the preferred method for creating
scatter plots using text markers.
• Starting with the third maintenance release of SAS 9.4, SUBPIXEL= specifies
whether subpixel rendering is used for image output when the scatter plot is
rendered.
SCATTERPLOTMATRIX:
• DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
xxvi Graph Template Language

• DATALABELSPLITCHARDROP= specifies whether the split characters are

included in the data labels.
• DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
• DATASKIN= enhances the visual appearance of the plot markers.
• ELLIPSE= now supports the CLIP= option, which specifies whether clipped
confidence ellipses are included in the plot.
• LABELSTRIP= specifies whether leading and trailing blanks are stripped from
marker characters or fixed-position data labels before they are displayed in the plot.
• MARKERATTRS= now supports transparency.
• MARKERCHARACTERPOSITION= specifies the justification of the marker
characters.
• MATRIXTYPE= specifies whether to display the full matrix, or just the upper or
lower triangle of the matrix.
• TIP= now supports NONE, which suppresses data tips from the plot.
• Starting with the second maintenance release of SAS 9.4, the
MARKERCOLORGRADIENT= option is replaced with the COLORRESPONSE=
option and is considered deprecated. The new option is more consistent with the
other plot statements. The MARKERCOLORGRADIENT= option is still honored,
but the COLORRESPONSE= option is the preferred option.
• Starting with the third maintenance release of SAS 9.4, SUBPIXEL= specifies
whether subpixel rendering is used for image output when the scatter plots are
rendered.
SERIESPLOT:
• CLUSTERAXIS= specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
• DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
• CURVELABELSPLIT= specifies whether to split the series line label at the
specified split characters.
• CURVELABELSPLITCHAR= specifies one or more characters on which the series
line label can be split if needed.
• CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the series line label text.
• CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the series line label block.
• DATASKIN= enhances the visual appearance of the series plot lines.
• FILLEDOUTLINEDMARKERS= specifies whether markers are drawn with both
fills and outlines.
Plot Statement Enhancements xxvii

• MARKERATTRS= now supports transparency.

• MARKERFILLATTRS= specifies the appearance of the filled markers.
• MARKEROUTLINEATTRS= specifies the appearance of the marker outlines.
• TIP= now supports NONE, which suppresses data tips and URLs from the plot.
• Starting with the second maintenance release of SAS 9.4, the following
enhancements are valid:
• The GROUPORDER= option supports REVERSEDATA, which orders the
groups within a category in the reverse group-column data order.
• LINECOLORGROUP= specifies a column that determines the line colors for a
grouped plot independently of the GROUP= column.
• LINEPATTERNGROUP= specifies a column that determines the line patterns for
a grouped plot independently of the GROUP= column.
• MARKERCOLORGROUP= specifies a column that determines the marker
colors for a grouped plot independently of the GROUP= column.
• MARKERSYMBOLGROUP= specifies a column that determines the marker
symbols for a grouped plot independently of the GROUP= column.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• ARROWHEADPOSITION= specifies a position for arrowheads.
• ARROWHEADSCALE= specifies an arrowhead scale factor based on the
thickness of the arrow line.
• ARROWHEADSHAPE= specifies a shape for arrowheads.
• COLORMODEL= specifies a color ramp that is to be used with the
COLORRESPONSE= option.
• COLORRESPONSE= specifies the column or range attribute variable that is
used to map the line and marker colors.
• LINETHICKNESSMAX= specifies the maximum line thickness when a
response variable is used to determine the line thickness.
• LINETHICKNESSMAXRESPONSE= specifies the response value that
corresponds to the maximum line thickness.
• LINETHICKNESSMIN= specifies the minimum line thickness when a response
variable is used to determine the line thickness.
• LINETHICKNESSRESPONSE= specifies a response column or range attribute
variable that is used to map a line thickness to each group value.
• SPLINEPOINTS= specifies a multiplier to apply to the time interval that is in
effect for the INTERVAL= axis option.
• SPLINETYPE= specifies the type of spline interpolation that is used to draw the
series line.
• The LINECOLORGROUP=, LINEPATTERNGROUP=,
MARKERCOLORGROUP=, and MARKERSYMBOLGROUP= options now
support a discrete attribute map variable.
STEPPLOT:
• CLUSTERAXIS= specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
xxviii Graph Template Language

• DATALABELSPLIT= specifies whether to split the data labels at specified split

characters.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
• DATASKIN= enhances the appearance of the step plot lines.
• ERRORBARCAPSHAPE= specifies whether the error bars have a serif cap. Starting
with the second maintenance release of SAS 9.4, this option defaults to style
reference GraphError:CapStyle.
• CURVELABELSPLIT= specifies whether to split the step line label at the specified
split characters.
• CURVELABELSPLITCHAR= specifies one or more characters on which the step
line label can be split if needed.
• CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the step line label text.
• CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the step line label block.
• FILLEDOUTLINEDMARKERS= specifies whether markers are drawn with both
fills and outlines.
• MARKERATTRS= now supports transparency.
• MARKERFILLATTRS= specifies the appearance of the filled markers.
• MARKEROUTLINEATTRS= specifies the appearance of the marker outlines.
• TIP= now supports NONE, which suppresses data tips and URLs from the plot.
• Starting with the second maintenance release of SAS 9.4, the GROUPORDER=
option supports REVERSEDATA, which orders the groups within a category in the
reverse group-column data order.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• ARROWHEADPOSITION= specifies a position for arrowheads.
• ARROWHEADSCALE= specifies an arrowhead scale factor based on the
thickness of the arrow line.
• ARROWHEADSHAPE= specifies a shape for arrowheads.
• LINETHICKNESSMAX= specifies the maximum line thickness when a
response variable is used to determine the line thickness.
• LINETHICKNESSMAXRESPONSE= specifies the response value that
corresponds to the maximum line thickness.
• LINETHICKNESSMIN= specifies the minimum line thickness when a response
variable is used to determine the line thickness.
• LINETHICKNESSRESPONSE= specifies a response column or range attribute
variable that is used to map a line thickness to each group value.
SURFACEPLOTPARM:
Plot Statement Enhancements xxix

• Starting with the second maintenance release of SAS 9.4, the

SURFACECOLORGRADIENT= option is replaced with the COLORRESPONSE=
option and is considered deprecated. The new option is more consistent with the
other plot statements. The SURFACECOLORGRADIENT= option is still honored,
but the COLORRESPONSE= option is the preferred option.
TEXTPLOT:
Starting with the third maintenance release of SAS 9.4, the following experimental
options are available:
• OUTFILE= specifies a file for storing information about the text bounding-box for
each text value in the column specified in the OUTID= option.
• OUTID= specifies a column that contains text values to write to the file specified in
the OUTFILE= option.
VECTORPLOT:
• DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
• DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
• DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
• DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
• DATASKIN= enhances the appearance of the vector plot lines.
• TIP= now supports NONE, which suppresses data tips from the plot.
• Starting with the third maintenance release of SAS 9.4, the following enhancements
are valid:
• COLORMODEL= specifies a color ramp that is used with the
COLORRESPONSE= option.
• COLORRESPONSE= specifies the column or range attribute variable that is
used to map the line colors.
• LINETHICKNESSMAX= specifies the maximum line thickness when a
response variable is used to determine the line thickness.
• LINETHICKNESSMAXRESPONSE= specifies the response value that
corresponds to the maximum line thickness.
• LINETHICKNESSMIN= specifies the minimum line thickness when a response
variable is used to determine the line thickness.
• LINETHICKNESSRESPONSE= specifies a response column or range attribute
variable that is used to map a line thickness to each group value.
WATERFALLCHART:
• BARLABELFITPOLICY= specifies a policy for avoiding collisions among the bar
labels when bar labels are displayed.
• BASELINEATTRS= specifies the appearance of the baseline. This option enables
you to suppress the baseline by setting the line thickness to 0.
• TIP= now supports NONE, which suppresses data tips and URLs from the plot.
xxx Graph Template Language

• Starting with the third maintenance release of SAS 9.4, COLORSTAT= specifies the
statistic to use for computing the response colors.

Text Statement Enhancements

ENTRY:
• Starting with the first maintenance release of SAS 9.4, the 512-character limit on the
length of the entry text is removed.
ENTRYFOOTNOTE:
• HALIGNCENTER= specifies whether the footnote is centered automatically by the
system or is always centered in the graph area.
• Starting with the first maintenance release of SAS 9.4, the 512-character limit on the
length of the footnote text is removed.
ENTRYTITLE:
• HALIGNCENTER= specifies whether the title is centered automatically by the
system or is always centered in the graph area.
• Starting with the first maintenance release of SAS 9.4, the 512-character limit on the
length of the title text is removed.

Draw Statement Enhancements

Starting with the first maintenance release of SAS 9.4, for the BEGINPOLYGON,
BEGINPOLYLINE, DRAWTEXT, DRAWLINE, DRAWARROW,
DRAWRECTANGLE, DRAWIMAGE, and DRAWOVAL statements, URL= specifies
an HTML page that is displayed when the output of this draw statement is selected.

Axis Statement Enhancements

LAYOUT OVERLAY:
• LABELFITPOLICY= specifies a policy for fitting axis labels in the available space.
• LABELPOSITION= specifies the position of the axis label.
• LABELSPLITCHAR= specifies one or more characters on which the axis labels can
be split if needed.
• LABELSPLITCHARDROP= specifies whether the split characters are included in
the axis labels.
• LABELSPLITJUSTIFY= specifies the justification of the strings that are inside the
axis label blocks.
• TICKVALUEHALIGN= specifies the horizontal alignment for all of the tick values
that are displayed on the Y and Y2 axes.
• TICKVALUEVALIGN= specifies the vertical alignment for all of the tick values that
are displayed on the X and X2 axes.
• Starting with the third maintenance release of SAS 9.4, LINEEXTENT= specifies
the extent of the axis line.
• DISCRETEOPTS= supports the following new features for discrete axes:
Axis Statement Enhancements xxxi

• TICKDISPLAYLIST= specifies the text that is displayed for the tick values that
are defined in the TICKVALUELIST= option.
• TICKVALUEFITPOLICY= supports new fit policies: ROTATE,
ROTATEALWAYS, ROTATELAWAYSDROP, SPLIT, SPLITALWAYS,
SPLITALWAYSTHIN, SPLITTHIN, and SPLITROTATE.
• TICKVALUELIST= specifies the list of tick values that are displayed on the
axis.
• TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
• TICKVALUESPLITCHAR= specifies a list of characters on which the tick
values can be split if needed.
• TICKVALUESPLITJUSTIFY= specifies justification of the strings that are
inside the tick value block.
• TICKVALUESPLITCHARDROP= specifies whether the split characters are
included in the displayed tick values.
• Starting with the third maintenance release of SAS 9.4, TICKVALUEFORMAT=
specifies how to format the values for major tick marks.
• LINEAROPTS= supports the following new features for linear axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKCOUNT= specifies the number of minor ticks that are displayed on
the axis.
• MINORTICKS= specifies whether the minor tick marks are displayed on the
axis.
• TICKDISPLAYLIST= specifies the text that is displayed for the tick values that
are defined in the TICKVALUELIST= option.
• TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
• TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which
enables you to specify the type of scale that you want to extract.
• TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
• Starting with the first maintenance release of SAS 9.4, INCLUDERANGES=
specifies the ranges for a broken axis.
• Starting with the second maintenance release of SAS 9.4:
• The MINORGRID= option defaults to style reference
GraphMinorGridLines:DisplayOpts.
• The MINORGRIDATTRS= option defaults to style element
GraphMinorGridLines in order to visually contrast the major and minor grid
lines.
• The MINORTICKCOUNT= option for linear axes defaults to one minor tick
and two intervals.
• LOGOPTS= supports the following new features for log axes:
xxxii Graph Template Language

• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKCOUNT= specifies the number of minor ticks that are displayed on
the axis.
• TICKVALUELIST= specifies the tick values for a log axis as a space-separated
list.
• TICKVALUEPRIORITY= specifies whether the TICKVALUELIST=
specification can extend the axis data range.
• VALUETYPES= specifies how the VIEWMIN=, VIEWMAX=, and
TICKVALUELIST= option values are interpreted.
• Starting with the second maintenance release of SAS 9.4:
• The MINORGRID= option defaults to style reference
GraphMinorGridLines:DisplayOpts.
• The MINORGRIDATTRS= option defaults to style element
GraphMinorGridLines in order to visually contrast the major and minor grid
lines.
• Starting with the third maintenance release of SAS 9.4, TICKVALUEFORMAT=
specifies how to format the values for major tick marks.
• TIMEOPTS= supports the following new features for time axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKINTERVAL= specifies the time interval between minor ticks.
• TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
• TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
• Starting with the first maintenance release of SAS 9.4, INCLUDERANGES=
specifies the ranges for a broken axis.
• Starting with the second maintenance release of SAS 9.4:
• The MINORGRID= option defaults to style reference
GraphMinorGridLines:DisplayOpts.
• The MINORGRIDATTRS= option defaults to style element
GraphMinorGridLines in order to visually contrast the major and minor grid
lines.
• Starting with the third maintenance release of SAS 9.4,
INTERVALMULTIPLIER= specifies a multiplier to apply to the time interval
that is in effect for the axis.
LAYOUT OVERLAY3D:
• LINEAROPTS= supports the following new features for linear axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
Axis Statement Enhancements xxxiii

• MINORTICKCOUNT= specifies the number of minor ticks that are displayed on

the axis.
• MINORTICKS= specifies whether the minor tick marks are displayed on the
axis.
• TICKDISPLAYLIST= specifies the text that is displayed for the tick values that
are defined in the TICKVALUELIST= option.
• TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which
enables you to specify the type of scale that you want to extract.
• TICKVALUEPRIORITY= specifies whether the TICKVALUELIST=
specification can extend the axis data range.
• Starting with the second maintenance release of SAS 9.4, the
MINORGRIDATTRS= option defaults to style element GraphMinorGridLines in
order to visually contrast the major and minor grid lines.
• TIMEOPTS= supports the following new features for time axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKINTERVAL= specifies the time interval between minor ticks.
• Starting with the second maintenance release of SAS 9.4, the
MINORGRIDATTRS= option defaults to style element GraphMinorGridLines in
order to visually contrast the major and minor grid lines.
• Starting with the third maintenance release of SAS 9.4,
INTERVALMULTIPLIER= specifies a multiplier to apply to the time interval
that is in effect for the axis.
LAYOUT OVERLAYEQUATED:
• MINORGRID= specifies whether grid lines are displayed at the minor tick values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKCOUNT= specifies the number of minor ticks that are displayed on the
axis.
• MINORTICKS= specifies whether the minor tick marks are displayed on the axis.
• TICKVALUEFITPOLICY= now supports policies NONE and ROTATEALWAYS.
• TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which enables
you to specify the type of scale that you want to extract.
• VIEWMAX= specifies the maximum data value to include in the display (the value
might be adjusted by the threshold calculation).
• VIEWMIN= specifies the minimum data value to include in the display (the value
might be adjusted by the threshold calculation).
• Starting with the third maintenance release of SAS 9.4, LINEEXTENT= specifies
the extent of the axis lines.
LAYOUT LATTICE:
• LABELFITPOLICY= specifies a policy for fitting axis labels in the available space.
• LABELPOSITION= specifies the position of the axis label.
xxxiv Graph Template Language

• LABELSPLITCHAR= specifies one or more characters on which the axis labels can
be split if needed.
• LABELSPLITCHARDROP= specifies whether the split characters are included in
the axis labels.
• LABELSPLITJUSTIFY= specifies the justification of the strings that are inside the
axis label blocks.
• TICKVALUEHALIGN= specifies the horizontal alignment for all of the tick values
that are displayed on the Y and Y2 axes.
• TICKVALUEVALIGN= specifies the vertical alignment for all of the tick values that
are displayed on the X and X2 axes.
• DISCRETEOPTS= supports the following new features for discrete axes:
• TICKDISPLAYLIST= specifies the text that is displayed for the tick values that
are defined in the TICKVALUELIST= option.
• TICKVALUEFITPOLICY= supports new fit policies: ROTATE,
ROTATEALWAYS, ROTATEALWAYSDROP, SPLIT, SPLITALWAYS,
SPLITALWAYSTHIN, and SPLITTHIN.
• TICKVALUELIST= specifies the list of tick values that are displayed on the
axis.
• TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
• TICKVALUESPLITCHAR= specifies a list of characters on which the tick
values can be split if needed.
• TICKVALUESPLITJUSTIFY= specifies justification of the strings that are
inside the tick value block.
• TICKVALUESPLITCHARDROP= specifies whether the split characters are
included in the displayed tick values.
• Starting with the third maintenance release of SAS 9.4, TICKVALUEFORMAT=
specifies how to format the values for major tick marks.
• LINEAROPTS= supports the following new features for linear axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKCOUNT= specifies the number of minor ticks that are displayed on
the axis.
• MINORTICKS= specifies whether the minor tick marks are displayed on the
axis.
• TICKDISPLAYLIST= specifies the text that is displayed for the tick values that
are defined in the TICKVALUELIST= option.
• TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
• TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which
enables you to specify the type of scale that you want to extract.
• TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
Axis Statement Enhancements xxxv

• Starting with the third maintenance release of SAS 9.4, TICKVALUEFORMAT=

specifies how to format the values for major tick marks.
• LOGOPTS= supports the following new features for log axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKCOUNT= specifies the number of minor ticks that are displayed on
the axis.
• TICKVALUELIST= specifies the tick values for a log axis as a space-separated
list.
• TICKVALUEPRIORITY= specifies whether the TICKVALUELIST=
specification can extend the axis data range.
• VALUETYPES= specifies how the VIEWMIN=, VIEWMAX=, and
TICKVALUELIST= option values are interpreted.
• Starting with the third maintenance release of SAS 9.4, TICKVALUEFORMAT=
specifies how to format the values for major tick marks.
• TIMEOPTS= supports the following new features for time axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKINTERVAL= specifies the time interval between minor ticks.
• TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
• TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
• Starting with the third maintenance release of SAS 9.4:
• TICKVALUEFORMAT= specifies how to format the values for major tick
marks.
• INTERVALMULTIPLIER= specifies a multiplier to apply to the time interval
that is in effect for the axis.
LAYOUT DATALATTICE and LAYOUT DATAPANEL:
• LABELFITPOLICY= specifies a policy for fitting axis labels in the available space.
• LABELPOSITION= specifies the position of the axis label.
• LABELSPLITCHAR= specifies one or more characters on which the axis labels can
be split if needed.
• LABELSPLITCHARDROP= specifies whether the split characters are included in
the displayed axis labels.
• LABELSPLITJUSTIFY= specifies the justification of the strings that are inside the
axis label blocks.
• TICKVALUEHALIGN= specifies the horizontal alignment for all of the tick values
that are displayed on the Y and Y2 axes.
• TICKVALUEVALIGN= specifies the vertical alignment for all of the tick values that
are displayed on the X and X2 axes.
xxxvi Graph Template Language

• DISCRETEOPTS= supports the following new features for discrete axes:

• TICKDISPLAYLIST= specifies the text that is displayed for the tick values that
are defined in the TICKVALUELIST= option.
• TICKVALUEFITPOLICY= supports new fit policies: ROTATE,
ROTATEALWAYS, ROTATEALWAYSDROP, SPLIT, SPLITALWAYS,
SPLITALWAYSTHIN, and SPLITTHIN.
• TICKVALUELIST= specifies the list of tick values that are displayed on the
axis.
• TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
• TICKVALUESPLITCHAR= specifies a list of characters on which the tick
values can be split if needed.
• TICKVALUESPLITJUSTIFY= specifies justification of the strings that are
inside the tick value block.
• TICKVALUESPLITCHARDROP= specifies whether the split characters are
included in the displayed tick values.
• Starting with the third maintenance release of SAS 9.4, TICKVALUEFORMAT=
specifies how to format the values for major tick marks.
• LINEAROPTS= supports the following new features for linear axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKCOUNT= specifies the number of minor ticks that are displayed on
the axis.
• MINORTICKS= specifies whether the minor tick marks are displayed on the
axis.
• TICKDISPLAYLIST= specifies the text that is displayed for the tick values that
are defined in the TICKVALUELIST= option.
• TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
• TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which
enables you to specify the type of scale that you want to extract.
• TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
• Starting with the second maintenance release of SAS 9.4, the
MINORGRIDATTRS= option defaults to style element GraphMinorGridLines in
order to visually contrast the major and minor grid lines.
• LOGOPTS= supports the following new features for log axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKCOUNT= specifies the number of minor ticks that are displayed on
the axis.
Discrete Attribute Map Enhancements xxxvii

• TICKVALUELIST= specifies the tick values for a log axis as a space-separated

list.
• TICKVALUEPRIORITY= specifies whether the TICKVALUELIST=
specification can extend the axis data range.
• VALUETYPES= specifies how the VIEWMIN=, VIEWMAX=, and
TICKVALUELIST= option values are interpreted.
• Starting with the second maintenance release of SAS 9.4, the
MINORGRIDATTRS= option defaults to style element GraphMinorGridLines in
order to visually contrast the major and minor grid lines.
• Starting with the third maintenance release of SAS 9.4, TICKVALUEFORMAT=
specifies how to format the values for major tick marks.
• TIMEOPTS= supports the following new features for time axes:
• MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
• MINORGRIDATTRS= specifies the attributes of the minor grid lines.
• MINORTICKINTERVAL= specifies the time interval between minor ticks.
• TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
• TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
• Starting with the second maintenance release of SAS 9.4, the
MINORGRIDATTRS= option defaults to style element GraphMinorGridLines in
order to visually contrast the major and minor grid lines.
• Starting with the third maintenance release of SAS 9.4,
INTERVALMULTIPLIER= specifies a multiplier to apply to the time interval
that is in effect for the axis.

Marker Definition Statement Enhancements

SYMBOLCHAR:
Starting with the third maintenance release of SAS 9.4, the following enhancements are
valid:
• HOFFSET= and VOFFSET= now move the marker character within the marker
character bounding box. The bounding box position remains centered on the data
point.
• Offsets applied to the marker character with HOFFSET= and VOFFSET= are also
applied to the marker character that is displayed in the legend.

Discrete Attribute Map Enhancements

The following enhancements have been made to discrete attribute maps:
VALUE statement:
• TEXTATTRS= specifies the text attributes to use when an attribute map is applied to
text in a plot.
xxxviii Graph Template Language

Starting with the first maintenance release of SAS 9.4, the following enhancements are
valid:
• You can now specify the attribute mapping information for a discrete attribute map in
a SAS data set. You now have an alternative to coding your mapping information in
a DISCRETEATTRMAP block in your template. For more information, see SAS
ODS Graphics: Procedures Guide.
• DISCRETEATTRMAP statement:
• DISCRETELEGENDENTRYPOLICY= specifies whether the items that the
associated plot contributes to a discrete legend are items that appear only in the
data or items that are defined only in the attribute map.
• DISCRETEATTRVAR statement:
• ATTRMAP= now accepts a name that is specified in the ID column in a discrete
attribute map data set. This feature enables you to create a discrete attribute map
variable for a discrete attribute map that is defined in a SAS data set. To resolve
the attribute map name in that case, you must specify the name of the attribute
map data set in the DATTRMAP= option in the SGRENDER statement that
renders the graph. For information about discrete attribute map data sets and the
SGRENDER statement DATTRMAP= option, see SAS ODS Graphics:
Procedures Guide.
• VALUE statement:
• FILLATTRS= now supports the TRANSPARENCY= fill option.
• LINEATTRS= now supports the THICKNESS= line option.
• MARKERATTRS= now supports the SIZE=, TRANSPARENCY=, and
WEIGHT= marker options.

Documentation Enhancements

In the second maintenance release of SAS 9.4, SAS Graph Template Language: User's
Guide is reorganized to make it easier to find information about how to use the Graph
Template Language.
1

Part 1

Fundamentals

Chapter 1
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2
3

Chapter 1
Overview

Graph Template Language (GTL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

GTL and the Output Delivery System (ODS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
A Quick Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Template Compilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Run-Time Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Basic Anatomy of an ODS Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Graphical Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Legends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Flexible Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Expressions and Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Dynamics and Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Conditional Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
ODS GRAPHICS Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
ODS Styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
About the Examples in This Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Examples and Resources on the Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Graph Template Language (GTL)

GTL and the Output Delivery System (ODS)

The SAS Graph Template Language (GTL) is an extension to the Output Delivery
System (ODS) that enables you to create sophisticated graphics. For example, using the
GTL, you can generate Model-Fit plots, Distribution Plots, Comparative plots,
Prediction Plots, and more.
The graphics produced by the GTL are generated by template definitions that control the
graph format and appearance and specify the variable roles to represent in the graph
display. The graph can then be rendered by associating the templates with a data source.
4 Chapter 1 • Overview

• The GTL templates are defined with PROC TEMPLATE. The GTL includes
conditional statements that can be used to determine what graph features are
rendered. It also includes layout statements that specify the arrangement of graph
features, plot statements that request specific plot types (such as histograms and
scatter plots), and text and legend statements that specify titles, footnotes, legends,
and other text-based graph elements.
• The GTL templates are rendered using the SGRENDER procedure, which specifies a
data source that contains appropriate data values and the template to use for
rendering the graph.
• You can also modify predefined GTL templates that the SAS System delivers for use
on the SAS statistical procedures. For information about modifying existing
templates, refer to SAS/STAT user’s guide.
This manual provides a complete reference to the Graph Template Language. For
detailed usage information, consult the SAS Graph Template Language: User's Guide.
Note: If you are also a SAS/GRAPH user, then you might want to consult the SAS
Graph Template Language: User's Guide to learn about some of the distinctions
between ODS Graphics and SAS/GRAPH.

A Quick Example
The data set Sashelp.Class is delivered with the SAS System. It includes data columns
named Height and Weight, which store height and weight measures for a small sample of
subjects. The Graph Template Language can be used to generate a histogram that shows
the distribution of weight recorded in that data set:

The following SAS program produces the graph:

proc template;
define statgraph histogram;
begingraph;
layout overlay;
Graph Template Language (GTL) 5

histogram weight;
endlayout;
endgraph;
end;
run;

ods graphics / width=450px;

proc sgrender data=sashelp.class
template=histogram;
run;

• The DEFINE STATGRAPH statement on PROC TEMPLATE opens a definition

block for defining a graphics template named HISTOGRAM. The HISTOGRAM
template is stored in the template folder (also called the “template store,” by default
located in Sasuser.Templat).
• The template definition for HISTOGRAM specifies two GTL statements within a
BEGINGRAPH/ENDGRAPH block: LAYOUT OVERLAY and HISTOGRAM.
• The LAYOUT OVERLAY statement is one of the most fundamental layout
statements. It can overlay the results of one or more plot statements, each of which
shares the same plot area, axes, and legends. The layout in this example specifies
only a single element: a HISTOGRAM with bars showing the distribution of
observations of the data column named Weight.
• The ENDLAYOUT statement ends the layout block, the ENDGRAPH statement
ends the graph definition, and the END statement ends the template definition.
• The ODS GRAPHICS statement uses the WIDTH= option to set a width for the
output graph. Because the HEIGHT= option is not specified, GTL manages the
graph’s aspect ratio and set an appropriate height.
• The DATA= option on PROC SGRENDER specifies Sashelp.Class as the data
source for the graph. TEMPLATE= specifies HISTOGRAM as the template
definition to use for rendering the graph.

Template Compilation
A GTL template describes the structure and appearance of a graph to be produced,
similar to how a TABLE template describes the organization and content of a table.
All templates are stored, compiled programs. The following source program produces a
simple GTL template named SCATTER:
proc template;
define statgraph scatter;
begingraph;
layout overlay;
scatterplot x=height y=weight;
endlayout;
endgraph;
end;
run;

When this code is submitted, the statement keywords and options are parsed, just as with
any other procedure. If no syntax error is detected, then an output template named
SCATTER is created and stored in the default template folder Sasuser.Templat. No graph
is produced. Note the following:
6 Chapter 1 • Overview

• Any required arguments in the template must be specified. In this example, X= and
Y= in the SCATTERPLOT statement must specify variables for the analysis, but no
checking for the existence of these variables is done at compile time. (Unlike other
SAS procedures, PROC TEMPLATE does not perform a compile and then run
sequence, which includes variable validation.)
• No reference to an input data set appears in the template.

Run-Time Actions
To produce a graph, a GTL template must be bound to a data source using the
SGRENDER procedure. The following example uses SGRENDER to bind the
SCATTER template to the SAS data set Sashelp.Class, which is delivered with the SAS
system:
proc sgrender data=sashelp.class
template=scatter;
run;

Generally, an ODS data object is constructed by comparing the template references to

column names with columns that exist in the current data set. In the current example,
Sashelp.Class contains columns named Height and Weight. Because these column names
match the columns that are named in template SCATTER, columns Height and Weight
are added to the data object. The other columns in Sashelp.Class are ignored. (It is
possible for a template to define new computed columns based on existing columns.)
After all the observations have been read, the data object and template definition are
passed to a graph renderer, which produces the graph image. The image is then
automatically integrated into the ODS destination. The visual properties of the graph are
determined by the ODS style that is in effect.
Note: Template SCATTER is a restrictive definition: it can create a plot only with
columns named Height and Weight. A GTL template can be made more flexible by
introducing dynamics or macro variables that supply variables and other information
at run time. For more information, see “Flexible Templates” on page 12.
Basic Anatomy of an ODS Graph 7

Basic Anatomy of an ODS Graph

The GTL is flexible and able to produce many different types of graphs with varying
layout features. The following figure shows the basic anatomy of an ODS graph:

Graph
the output produced from all of the statements that are nested in a BEGINGRAPH
statement block. The graph comprises all of the graphics elements in the template
definition.
Title Area
area for one or more titles. This area is always displayed above all cells in the graph.
Footnote Area
area for one or more footnotes. This area is always displayed below all cells in the
graph.
Cell
refers collectively to the area containing the plot areas. In this diagram, there are two
cells, each of which contains two axes for the plot area. A cell can also contain
descriptive text and legends. Graphs are often described as single-cell or multi-cell.
Plot Area
the display area for plot-statement results. This area is bounded by the axes (when
present) and can also contain data labels and other text that annotates the graph.
Axis
refers collectively to the axis line, the major and minor tick marks, the major tick
values, and the axis label.
8 Chapter 1 • Overview

Plots
refers collectively to all plot statements that can be overlaid in the plot area. This
includes graphical items such as fit lines, scatter plots, reference lines, and many
others.
Legend
refers collectively to one or more legend entries, each made up of a graphical value
and a text label. The legend can also have a title and border. Legends can also
display a color ramp corresponding to a continuous response range.

Graphical Layouts
[ˌhaɪəˈr :rkɪkl]:分层
One of most powerful features of the GTL is the syntax built around hierarchical
statement blocks called “layouts.” The outermost layout block determines
• The overall organization of the graph—whether it uses a single-cell or a multi-cell
display.
• What statements are allowed in the block. Generally, layout blocks can contain plots,
lines of text, a legend, or even another layout.
• How the contained statements interact.

Table 1.1 Outermost Layouts in GTL

Layout Description

OVERLAY General purpose layout for displaying 2-D plots in a single-cell.

OVERLAY3D Layout for displaying 3-D plots in a single-cell.

OVERLAYEQUATED Specialized OVERLAY with equated axes.

等同

REGION General purpose layout for displaying single-cell graphs that does
not use axes.

GRIDDED Basic grid of plots. All cells are independent.

LATTICE Advanced multi-cell layout. Axes can be shared across columns

or rows and be external to grid. Many grid labeling and alignment
features.

DATALATTICE Generates a classification panel from the values of 1 or 2

classifiers.

DATAPANEL Generates a classification panel from the values of n classifiers.

GLOBALLEGEND Specialized layout for creating a compound legend that contains

multiple discrete legends.

For example, the following graph is a two-cell graph produced using the LAYOUT
LATTICE statement as the outermost template in the layout.
Graphical Layouts 9

The LAYOUT LATTICE statement is typically used to create a multi-cell layout of plots
that are aligned across columns and rows. In the following template, which produced the
graph, plot statements are specified within nested LAYOUT OVERLAY statements.
Thus, the LATTICE automatically aligns the plot areas and tick display areas in the
plots. The LATTICE layout is a good layout to choose when you want to compare the
results of related plots.
proc template;
define statgraph lattice;
begingraph;
entrytitle "Car Performance Profile";
layout lattice / border=true pad=10 opaque=true
rows=1 columns=2 columngutter=3;
layout overlay;
scatterplot x=horsepower y=mpg_city /
group=origin name="cars";
regressionPlot x=horsepower y=mpg_city / degree=2;
endlayout;

layout overlay;
scatterplot x=weight y=mpg_city / group=origin;
regressionPlot x=weight y=mpg_city / degree=2;
endlayout;

sidebar;
discretelegend "cars";
endsidebar;
endlayout;
endgraph;
end;
run;

For detailed information about each layout, see the chapter for that layout type.
10 Chapter 1 • Overview

Plots
The plots in the GTL are classified in different ways, depending on the context of the
discussion.
Within layout blocks, plots are often classified according to graphical dimension:
whether they are projected in two or three visual dimensions. Thus, plots in the GTL are
often referred to as 2-D or 3-D plots, based on their graphical dimensions, not their data
dimensions.
Relative to their input data, plots are classified according to the statements that calculate
summary statistics from raw input data, and those that use calculated statistics as input
parameters on the plot statement. Thus, many GTL plot statements have two versions:
BARCHART and BARCHARTPARM, HISTOGRAM and HISTOGRAMPARM, and so
on. The main distinction between such plots is the nature of the input data that they
accept:
• The “non-parm” version (for example, BARCHART) computes its values from raw,
unsummarized data. For example, a BARCHART computes the summary values it
needs for the bars in the chart. Such plots are often referred to as “computed plots.”
• The “parm” version (for example, BARCHARTPARM) does not summarize or
compute values from the input data but instead simply renders the input data it is
given. Thus, the input data must be pre-summarized, perhaps by a SAS procedure.
The “parm” version of plots, often referred to as “parameterized plots,” produce the
same result as the non-parm version. However, they do not perform the calculations
or data summarizations needed to achieve the result.
Chapter 5, “Key Concepts for Using Plots,” on page 175 discusses general concepts that
apply across plot types. For detailed information about a particular plot, see the chapter
for that plot.

Axes
The GTL uses various criteria to determine the displayed axis features for a graph.
Generally, axis features are based on the layout type, the order of plot statements in the
layout and the options specified on those statements, the use of “primary” and
“secondary” axes on the plots (when secondary axes are supported), the plot type, the
column(s) of data that contribute to defining the axis range, and the data formats for the
contributing data columns.
Depending on the layout type, 2-D plots can have up to four independent axes that can
be displayed: X, Y, X2, and Y2. The X and Y axes are considered the primary axes, and
the X2 and Y2 axes are considered the secondary axes. By default, the X2 and Y2 axes
are not displayed. When requested, the secondary axes can be displayed as copies of the
primary axes, or data can be mapped separately to them. The following figure identifies
the X, Y, X2, and Y2 axes.
Legends 11

All 3-D plots display the standard X, Y, and Z axes.

For more information about axis features in GTL, see Chapter 7, “Axis Features in
Layouts,” on page 875.

Legends
Many plot statements support a GROUP= option that partitions the data into unique
values, performs separate analysis, if necessary, and automatically assigns distinct visual
properties to each group value. The visual properties of group values are defined by the
style in effect.
Legends are not automatically displayed for plots with group values. Rather, an
appropriate legend statement must be added to the template to generate the desired
legend. In the following example, a legend is added to display markers and line patterns
that show the association between the group values from a scatter plot and corresponding
linear regression lines. The example shows the mechanism that GTL uses to associate a
legend with its corresponding plot(s): a name is assigned to each plot that must be
represented in the legend, and these names are then used as arguments for the legend
statement (in this case, MERGEDLEGEND).
proc template;
define statgraph scatterfit;
begingraph;
12 Chapter 1 • Overview

entrytitle "Linear Regression By Gender";

layout overlay;
scatterplot x=height y=weight / group=sex name="scat";
regressionplot x=height y=weight/ group=sex name="reg";
mergedlegend "scat" "reg" / border=true;
endlayout;
endgraph;
end;
run;

For more information about managing legends in GTL, see SAS Graph Template
Language: User's Guide.

Flexible Templates
特征 限制性的
Several features in the GTL can make template definitions less restrictive on input data
and more general in nature. These features enable a single compiled template to produce
many output variations.
变化

Expressions and Functions

In the GTL, expressions can be used to compute constants and data columns. The
expressions must be enclosed in an EVAL construct. Within the expression, you can use
DATA step functions, arithmetic operators, and other special functions supported by the
GTL. 算术
Expressions are also useful in text statements like ENTRY and ENTRYTITLE. Both of
these statements support rich text and have special text commands such as {SUP},
{SUB}, and {UNICODE}, which enable subscripting, superscripting, and Unicode
characters.
The following template shows how the ± symbol is included in the title line using its
十六进制hexadecimal Unicode value. Also, new data columns are computed for the upper and
lower error bars of the scatter plot, based on the input columns MeanWeight and
STDERR.
proc template;
define statgraph expression;
begingraph;
entrytitle "Errorbars show " {unicode "00B1"x} "2 SE";
Flexible Templates 13

layout overlay;
scatterplot x=age y=meanweight /
yerrorlower=eval(meanweight - 2*stderr)
yerrorupper=eval(meanweight + 2*stderr);
seriesplot x=age y=meanweight;
endlayout;
endgraph;
end;
run;

For more information about using expressions, see Chapter 21, “Expressions,” on page
1317. For more information about using functions, see Chapter 22, “Functions,” on page
1321.

Dynamics and Macro Variables

An extremely useful technique for generalizing templates is to define dynamics, macro
variables, or both. The dynamics and macro variables resolve when the template is
executed. The following PROC TEMPLATE statements can be used in a DEFINE
STATGRAPH block:

Template Statement Purpose Value supplied by...

DYNAMIC defines one or more dynamic either of the following:

variables
• DYNAMIC= suboption of
ODS= option of FILE
PRINT
• DYNAMIC statement of
PROC SGRENDER

MVAR defines one or more macro %LET or CALL SYMPUT( )

variables

NMVAR defines one or more macro %LET or CALL SYMPUT( )

variables that resolve to a
number or numbers

NOTES provides information about user-supplied text

the graph definition

The following example defines a template named DYNAMICS that can create a
histogram and density plot for any variable. It defines both macro variables and
dynamics for run-time substitution. No data-dependent information is hard coded in the
template.
Note: You can initialize macro variables with %LET statements and dynamics with
SGRENDER’s DYNAMIC statement.
proc template;
define statgraph dynamics;
mvar SYSDATE9 SCALE;
nmvar BINS;
dynamic VAR VARLABEL;
begingraph;
entrytitle "Histogram of " VAR;
14 Chapter 1 • Overview

entrytitle "with Normal Distribution";

layout overlay / xaxisopts=(label=VARLABEL);
histogram VAR / scale=SCALE nbins=BINS;
densityplot VAR / normal();
endlayout;
entryfootnote halign=right "Created: " SYSDATE9 /
textattrs=GraphValueText;
endgraph;
end;
run;

%let bins=6;
%let scale=count;
proc sgrender data=sashelp.class
template=dynamics;
dynamic var="Height" varlabel="Height in Inches";
run;

For more information about using dynamics and macro variables, see Chapter 20,
“Dynamics and Macro Variables,” on page 1313.

Conditional Logic
Using conditional logic, you can create templates that have multiple visual results or
output representations, depending on existing conditions. The evaluation of a logical
expression must generate one or more complete statements (not portions of statements).
All conditional logic uses one of the following constructs:

if (condition) if (condition)
statement(s); statement(s);
endif; else
statement(s);
endif;

In the IF statement, condition must be enclosed in parentheses. The condition can be any
standard SAS expression involving arithmetic, logical operators, comparison operators,
Boolean operators, or concatenation operators. The expression can also use SAS DATA
step functions. The expression resolves to a single numeric value, which is true or false.
Output 15

In the following example, a histogram is conditionally overlaid with a normal

distribution curve, a Kernel Density Estimate distribution curve, both, or neither:
proc template;
define statgraph conditional;
dynamic VAR VARLABEL BINS CURVE;
begingraph;
entrytitle "Histogram of " VAR;
layout overlay / xaxisopts=(label=VARLABEL);
histogram VAR / nbins=BINS;

if (upcase(CURVE) in ("ALL" "KERNEL"))

densityplot VAR / kernel() name="k"
legendlabel="Kernel"
lineattrs=(pattern=dash);
endif;

if (upcase(CURVE) in ("ALL" "NORMAL"))

densityplot VAR / normal() name="n"
legendlabel="Normal";
endif;

discretelegend "n" "k";

endlayout;
endgraph;
end;
run;

Note that the legend syntax does not have to be made conditional. At run time, each plot
name in the legend is checked. If the plot does not exist, then its name is removed from
the legend name list. If no names appear in the DISCRETELEGEND statement, then the
legend “drops out” and the histogram size is adjusted to fill the remaining space.
For more information about using conditional logic, see Chapter 23, “Conditional
Logic,” on page 1333.

Output
When using the GTL, you focus primarily on defining template definitions that produce
specific graphs and generate a particular output layout. Ultimately, you must also
customize the graphical environment to get the exact output that you desire. The ODS
GRAPHICS statement is available for customizing the graphical environment, and ODS
styles enable you to manage the output appearance.

ODS GRAPHICS Statement

The ODS GRAPHICS statement is used to modify the environment in which graphics
templates are executed. The ODS GRAPHICS statement is used to control
• whether ODS graphics is enabled
• the type and name of the image created
• the size of the image
• whether features such as scaling and anti-aliasing are used.
16 Chapter 1 • Overview

The following ODS GRAPHICS statement uses the HEIGHT= and WIDTH= options to
set an aspect ratio for the output image.
ods graphics on / height=175px width=200px;
proc sgrender data=sashelp.class
template=scatter;
run;
ods graphics off;

For more information about using the ODS GRAPHICS statement in GTL, see SAS
Graph Template Language: User's Guide. For a more complete discussion of the ODS
GRAPHICS statement, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.

ODS Styles
When any graphics template is executed, there is always an ODS style in effect that
governs the appearance of the output. The following ODS statement sends graphics
output to the RTF output destination using the LISTING style:
ods rtf style=listing;

ods graphics on / height=175px width=200px border=off;

proc sgrender data=sashelp.class
template=scatter;
run;
ods graphics off;

ods rtf close;

Support for ODS styles is highly integrated into GTL syntax. By default, the graphical
appearance features of most plot and text statements are mapped to corresponding style
elements and associated attributes. Because of this, your output tables and graphs always
have a reasonable overall appearance. Moreover, output for a given ODS destination has
a consistent look (for example, table colors and graph colors do not clash).
The following figures show how a graph’s appearance can be changed by using
references to style elements to set the graph’s appearance options. This technique
permits changes in graph appearance by style modification instead of graphical template
modification. The graphs in the figures are generated with the following GTL statement:
contourplotparm x=x y=y z=density /
contourtype=fill nhint=9
colormodel=ThreeColorRamp ;

The following style template shows the definition for the ThreeColorRamp style
element:
style ThreeColorRamp /
endcolor = GraphColors("gramp3cend")
neutralcolor = GraphColors("gramp3cneutral")
startcolor = GraphColors("gramp3cstart");
About the Examples in This Documentation 17

For more information about the use of ODS styles in GTL, see SAS Graph Template
Language: User's Guide. For a more complete discussion of ODS styles, see SAS Output
Delivery System: User's Guide.

About the Examples in This Documentation

The example programs that are shown in this document often provide all of the code that
you need to generate the graphs that are shown in the figures. We encourage you to copy
and paste the example code into your SAS session and generate the graphs for yourself.
The examples are written to be runnable in the SAS windowing environment and in SAS
Studio. Unless otherwise noted, the examples use the default ODS destination. In the
SAS windowing environment, the default ODS destination is ODS HTML. For
information about the default ODS output in SAS Studio, see “SAS Studio and ODS” in
SAS Output Delivery System: User's Guide. For information about using SAS Studio, see
SAS Studio: User's Guide.
If you generate the example graphs using an HTML destination, they are typically
rendered as 640 pixel by 480 pixel images using the HTMLBlue style. Because of size
limitations, the graphs in this document are not shown in their default size. They are
scaled down to meet the size requirements of our documentation production system.
When graphs that are produced with ODS graphics are reduced in size, several automatic
processes take place to optimize the appearance of the output. Among the differences
between default size graphs and smaller graphs are that the smaller graphs have scaled
18 Chapter 1 • Overview

down font sizes. Also, their numeric axes might display a reduced number of ticks and
tick values. Thus, the graphs that you generate from the example programs will not
always look identical to the graphs that are shown in the figures. However, both graphs
will accurately represent the data.
When you produce your own graphical output, you can change the graph size and
attributes, if needed. The SAS Graph Template Language: User's Guide explains how to
set fonts, DPI, anti-aliasing, and other features that contribute to producing professional-
looking graphics of any size in any output format.

Examples and Resources on the Web

The SAS website contains a large number of examples that can help you visualize and
code your graphs. The examples cover a range of SAS technologies including the ODS
Graphics procedures.
• Graphically Speaking is a blog by Sanjay Matange focused on using ODS Graphics
for data visualization in SAS. The blog covers topics related to the ODS Graphics
procedures, the SAS Graph Template Language, and the SAS ODS Graphics
Designer.
http://blogs.sas.com/content/graphicallyspeaking/
• The Graphics Samples Output Gallery is a collection of graphs organized by SAS
procedure. The graphs link to the source code in SAS Samples & Notes. The gallery
is maintained by SAS Technical Support.
http://support.sas.com/sassamples/graphgallery/index.html
• The Focus Area Graphics site provides a simple interface to business and analytical
graphs. The site is maintained by the SAS Data Visualization team.
http://support.sas.com/rnd/datavisualization/index.htm
• Samples & SAS Notes contains an abundance of searchable examples. You can
browse by topic, search for a particular note, search for a particular technology such
as SGPLOT, and conduct other searches.
http://support.sas.com/notes/index.html
In addition, you can share your questions, suggestions, and experiences related to
graphics on the SAS/GRAPH and ODS Graphics community site. See https://
communities.sas.com/community/support-communities/sas_graph_and_ods_graphics.
19

Part 2

Graph Block

Chapter 2
BEGINGRAPH Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
20
21

Chapter 2
BEGINGRAPH Statement

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
BEGINGRAPH Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Dictionary

BEGINGRAPH Statement
Defines the outermost container for a graph template that is defined with GTL-statements.
Requirements: All STATGRAPH template definitions must start with a BEGINGRAPH statement and
end with an ENDGRAPH statement.
The BEGINGRAPH block must contain one and only one layout block.
The layout block and its nested layouts, if any, must contain at least one plot.

Syntax
BEGINGRAPH </option(s)>;
<GTL-global-statements>
GTL-layout-block
<GTL-global-statements>
ENDGRAPH;

Summary of Optional Arguments

Appearance options
ATTRPRIORITY=AUTO | COLOR | NONE
specifies a priority for cycling the group attributes.
BACKGROUNDCOLOR=style-reference | color
specifies the color of the graph background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the graph.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the graph.
DATACOLORS=(color-list)
22 Chapter 2 • BEGINGRAPH Statement

specifies the list of fill colors that will replace the graph data colors from the
GraphData1–GraphDataN style elements.
DATACONTRASTCOLORS=(color-list)
specifies the list of contrast colors that will replace the graph data contrast
colors from the GraphData1–GraphDataN style elements.
DATALINEPATTERNS=(line-pattern-list)
specifies the list of line patterns that will replace the graph data line patterns
from the GraphData1–GraphDataN style elements.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of all plots in the template that support data
skins.
DATASYMBOLS=(marker-symbol-list)
specifies the list of marker symbols that will replace the graph data marker
symbols from the marker symbols that are defined in the GraphData1–
GraphDataN style elements.
DESIGNHEIGHT=DEFAULTDESIGNHEIGHT | dimension
specifies the design height of the graph.
DESIGNWIDTH=DEFAULTDESIGNWIDTH | dimension
specifies the design width of the graph.
DRAWSPACE= GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a global drawing space and drawing units for all of the draw
statements within this BEGINGRAPH block.
OPAQUE=TRUE | FALSE
specifies whether the graph background is opaque or transparent. 不透明或透明
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the graph border.
SUBPIXEL=AUTO | OFF | ON
specifies whether subpixel rendering is used for drawing smooth curved lines
or for spacing bars more precisely.

Axis options
AXISBREAKSYMBOL=BRACKET | NOTCH | SLANTEDLEFT |
SLANTEDRIGHT | SQUIGGLE | SPARK | Z
specifies a symbol to use on the axis lines to indicate a break in the axis.
AXISBREAKTYPE=FULL | AXIS
specifies whether the axis break is indicated in the full display or only on the
axis line.
AXISLINEEXTENT=FULL | DATA | number
specifies the extent of the axis line for all axes.
DISCRETEAXISOFFSETPAD=TRUE | FALSE
specifies whether additional padding is added to the minimum and maximum
axis offsets for discrete axes.

Label options
LABELPLACEMENT=AUTO | GREEDY | SA
specifies the label-placement algorithm to use for positioning labels in the
graphs.
SAPLACEMENTOPTS=(placement-options)
specifies the options for the label-placement algorithm when
LABELPLACEMENT=SA.
BEGINGRAPH Statement 23

Midpoint options
INCLUDEMISSINGDISCRETE=TRUE | FALSE
specifies whether missing values are displayed on a discrete axis.

Optional Arguments
ATTRPRIORITY=AUTO | COLOR | NONE
specifies a priority for cycling the group attributes.
AUTO
honors the current state of the attribute priority rotation pattern as specified in the
active style or in the ODS GRAPHICS statement.
COLOR
changes the current setting of attribute priority rotation pattern to the color-
priority pattern by cycling through the list of colors while holding the marker
symbol, line pattern, or fill pattern constant. When all of the colors are exhausted,
the marker symbol, line style, or fill pattern attribute increment to the next
element, and then the colors in the list are repeated. This pattern repeats as
needed.
NONE
changes the current setting of attribute priority rotation pattern to the default
pattern, which cycles progressively through the attribute lists.

Default The attribute priority pattern that is specified in the active style or in
the ODS GRAPHICS statement.

Interactions This option overrides the attribute priority rotation pattern that is
specified in the current style and the ATTRPRIORITY= option in the
ODS GRAPHICS statement.

The default lists of data colors, contrast colors, marker symbols, and
line patterns are set in the active style’s GraphData1–GraphDataN
elements.

The individual attributes in these lists can be overridden with the

BEGINGRAPH options DATACOLORS=,
DATACONTRASTCOLORS=, DATALINEPATTERNS=, and
DATASYMBOLS=.

The ATTRPRIORITY= option affects the cycling of the style

attributes for GROUP=, CYCLEATTRS=TRUE, and explicit style
references such as MARKERATTRS=GraphData2.

See “Attribute Rotation Patterns” in SAS Graph Template Language:

User's Guide

AXISBREAKSYMBOL=BRACKET | NOTCH | SLANTEDLEFT |

SLANTEDRIGHT | SQUIGGLE | SPARK | Z
specifies a symbol to use on the axis lines to indicate a break in the axis.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
24 Chapter 2 • BEGINGRAPH Statement

The following figure shows an example of each symbol on a horizontal linear axis for
ranges 1–4 and 6–10.

Default SQUIGGLE

Restriction This option applies to linear and time axes only.

Requirements The AXISBREAKTYPE= option must be set to AXIS for this

option to have any effect.

You must use the INCLUDERANGES= option to specify ranges for

the axis for this option to have any effect.

The DISPLAY= option for the axis must include the axis line for
this option to have any effect.

AXISBREAKTYPE=FULL | AXIS
specifies whether the axis break is indicated in the full display or only on the axis
line.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
The following figure shows an example of each type for ranges 50–52 and 56–73 on a
linear horizontal axis.
Figure 2.1 Axis Break Types FULL and AXIS

Default FULL

Requirements You must use the INCLUDERANGES= axis option to specify

ranges for the axis for this option to have any effect.

The axis line or the plot wall outline must be displayed for AXIS to
have any effect. Otherwise, FULL is used.

Interaction When AXIS is specified, if the secondary axis line or the plot wall
outline is displayed, then the axis break symbol is displayed on both
the primary and the secondary axis. Otherwise, the break symbol is
BEGINGRAPH Statement 25

displayed only on the primary axis, as shown in Figure 2.1 on page

24.

Notes The axis break indicators pass through inner margin areas.

No attempt is made to avoid collisions between the axis break

indicators and other graphical elements.

Tip When you use AXIS, use the AXISBREAKSYMBOL= option to

change the break symbol.

AXISLINEEXTENT=FULL | DATA | number

specifies the extent of the axis line for all axes.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
FULL
specifies axis lines that extend along the entire length of the axis.
DATA
specifies axis lines that extend through the data range from the minimum offset
to the maximum offset.
number
specifies, as a decimal proportion, how much the axis line extends from DATA
toward FULL. A value of 0 is equivalent to DATA, and a value of 1 is equivalent
to FULL.

Range 0–1

Tip A numeric value is useful for bar charts when DATA terminates the axis
line at the midpoint positions of the minimum and maximum bars. In
that case, you can specify a numeric value to lengthen the axis line so
that it extends to the full width of both bars.

The following figure shows a simple example of each value for the X and Y axis lines.
The light-blue dashed lines depict the minimum and maximum offsets that are set on the
axes.

Default FULL

Restriction This option is valid only in OVERLAY and OVERLAYEQUATED

layouts.

Tips The graph wall outline might appear to be an axis line. In that case, use
the WALLDISPLAY=NONE or WALLDISPLAY=(FILL) option in the
layout statement to suppress the wall outline.
26 Chapter 2 • BEGINGRAPH Statement

Use the LINEEXTENT= axis option to control the axis line extent on a
per-axis basis.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the graph background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction This option has no effect when OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the graph.

Default The ODS GRAPHICS statement BORDER= option setting, which is

TRUE by default.

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the graph.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element .

“Line Options” on page 1349 for available line-options.

DATACOLORS=(color-list)
specifies the list of fill colors that will replace the graph data colors from the
GraphData1–GraphDataN style elements.
(color-list)
a space-separated list of colors, enclosed in parentheses. You can use a style
attribute reference such as GraphData3:color, a color name, or an RGB, CMYK,
HLS, or HSV (HSB) color code to specify a color. The list can contain a mix of
style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

Example datacolors=(CXFF0000 green blue)

When this option is specified, the fill colors rotate through this color list rather than
through the colors that are defined in the GraphData1–GraphDataN style elements.
For information about the attribute rotation patterns, see “Attribute Rotation
Patterns” in SAS Graph Template Language: User's Guide.

Default The colors that are defined in the GraphData1–GraphDataN style

elements.
BEGINGRAPH Statement 27

Interaction Where applicable, the COLOR= suboption of the FILLATTRS= option

overrides the DATACOLORS= option.

DATACONTRASTCOLORS=(color-list)
specifies the list of contrast colors that will replace the graph data contrast colors
from the GraphData1–GraphDataN style elements.
(color-list)
a space-separated list of contrast colors, enclosed in parentheses. You can use a
style attribute reference such as GraphData3:color, a color name, or an RGB,
CMYK, HLS, or HSV (HSB) color code to specify a color. The list can contain a
mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

Example datacontrastcolors=(orange cyan #FF0000)

When this option is specified, the contrast colors cycle through this color list rather
than through the contrast colors that are defined in the GraphData1–GraphDataN
style elements. For information about the attribute rotation patterns, see “Attribute
Rotation Patterns” in SAS Graph Template Language: User's Guide.

Default The contrast colors that are defined in the GraphData1–GraphDataN

style elements.

Interaction Where applicable, the COLOR= suboption of the MARKERATTRS=

option or the LINEATTRS= option overrides the
DATACONTRASTCOLORS= option.

DATALINEPATTERNS=(line-pattern-list)
specifies the list of line patterns that will replace the graph data line patterns from the
GraphData1–GraphDataN style elements.
(line-pattern-list)
a space-separated list of line patterns, enclosed in parentheses. You can use a
style attribute reference such as GraphData3:lineStyle, a line pattern number, or a
line pattern name (where applicable) to specify a pattern. The list can contain a
mix of style attribute references, line pattern numbers, and line pattern names.

Requirement The list of line patterns must be enclosed in parentheses.

When this option is specified, the line patterns cycle through this line-pattern list
rather than through the line patterns that are defined in the GraphData1–GraphDataN
style elements. When the patterns in line-pattern-list are exhausted, the patterns
repeat.

Default The line patterns that are defined in the GraphData1–GraphDataN style
elements.

Interaction Where applicable, the PATTERN= suboption of the LINEATTRS=

option overrides the DATALINEPATTERNS= option.

Example datalinepatterns=(solid dash 16 26)

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of all plots in the template that support data skins.
The following plot statements support data skins:
28 Chapter 2 • BEGINGRAPH Statement

BARCHART HISTOGRAM SCATTERPLOT

BARCHARTPARM HISTOGRAMPARM SCATTERPLOTMATRIX
BOXPLOT LINECHART SERIESPLOT
BOXPLOTPARM NEEDLEPLOT STEPPLOT
BUBBLEPLOT PIECHART VECTORPLOT
DROPLINE POLYGONPLOT WATERFALLCHART
HIGHLOWPLOT REFERENCELINE

Default The GraphSkins:DataSkin style attribute, if it is specified in the current

style. If the current style does not specify the GraphSkins:DataSkin
style attribute, then the default is NONE.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interaction This option is overridden by the DATASKIN= option in the individual

plot statements.

Note Applying data skins to a graph that has a very large number of markers
can negatively impact performance.

DATASYMBOLS=(marker-symbol-list)
specifies the list of marker symbols that will replace the graph data marker symbols
from the marker symbols that are defined in the GraphData1–GraphDataN style
elements.
(marker-symbol-list)
a space-separated list of marker symbols, enclosed in parentheses. You can use a
style attribute reference such as GraphData5:markerSymbol or a marker symbol
name to specify a marker. The list can contain a mix of style attribute references
and marker symbol names.

Requirement The list of marker symbols must be enclosed in parentheses.

When this option is specified, the marker symbols cycle through this marker symbol
list rather than through the line patterns that are defined in the GraphData1–
GraphDataN style elements. When the symbols in marker-symbol-list are exhausted,
the symbols repeat.

Default The marker symbols that are defined in the GraphData1–GraphDataN

style elements.

Interaction Where applicable, the SYMBOL= suboption of the MARKERATTRS=

option overrides the DATASYMBOLS= option.

Example datasymbols=(circle square triangle star)

DESIGNHEIGHT=DEFAULTDESIGNHEIGHT | dimension
specifies the design height of the graph.
BEGINGRAPH Statement 29

Default DEFAULTDESIGNHEIGHT. This value is obtained from the SAS

Registry key ODS ð ODS Graphics ð Design Height when the
graph is rendered. The initial value of this registry key is 480px.

Restriction The minimum dimension value that you can set is 2 pixels. If a smaller
setting is specified, then the default design height is used.

Interactions The design height can be overridden at run time with a render height
that is specified with the HEIGHT= option in the ODS GRAPHICS
statement (external to the template). Also, the ODS destination
statement’s IMAGE_DPI= option can affect the height.

You can change the value of the Design Height registry key in the
SAS registry. However, doing so affects the design height of all
templates that do not include an explicit dimension for the design
height. You can also change the height setting in the graph style, but
doing so affects the height of all templates that use that style.

See “dimension” on page 1340

DESIGNWIDTH=DEFAULTDESIGNWIDTH | dimension
specifies the design width of the graph.

Default DEFAULTDESIGNWIDTH. This value is obtained from the SAS

Registry key ODS ð ODS Graphics ð Design Width when the
graph is rendered. The initial value of this registry key is 640px.

Restriction The minimum dimension value that you can set is 2 pixels. If a smaller
setting is specified, then the default design width is used.

Interactions The design width can be overridden at run time with a render width
that is specified with the WIDTH= option in the ODS GRAPHICS
statement (external to the template). Also, the ODS destination
statement’s IMAGE_DPI= option can affect the width.

You can change the value of the Design Width registry key in the SAS
registry. However, doing so affects the design width of all templates
that do not include an explicit dimension for the design width. You can
also change the width setting in the graph style, but doing so affects
the width of all templates that use that style.

See “dimension” on page 1340

DISCRETEAXISOFFSETPAD=TRUE | FALSE
specifies whether additional padding is added to the minimum and maximum axis
offsets for discrete axes. When set to TRUE, an additional 5 pixels of padding is
added to the minimum and maximum axis offsets.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default TRUE

Restriction This option applies only to discrete axes.

Tip This option is useful for heat maps when you want the heat map to
occupy the entire plot area. In that case, in addition to setting this
30 Chapter 2 • BEGINGRAPH Statement

option to FALSE, set OFFSETMIN= and OFFSETMAX= to 0 for the

discrete axes.

DRAWSPACE= GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a global drawing space and drawing units for all of the draw statements
within this BEGINGRAPH block.

Default LAYOUTPERCENT

Tip Individual draw statements within this BEGINGRAPH block can override
this global setting.

See “About the Drawing Space and Drawing Units” on page 1192

INCLUDEMISSINGDISCRETE=TRUE | FALSE
specifies whether missing values are displayed on a discrete axis.

Default FALSE

Interaction This option affects all charts and plots within the template.

See “boolean ” on page 1339 for other Boolean values that you can use.

LABELPLACEMENT=AUTO | GREEDY | SA
specifies the label-placement algorithm to use for positioning labels in the graphs.
The following labels are affected:
• data labels for needle plots, scatter plots, series plots, step plots, and vector plots
• vertex labels for line charts
• curve labels when the curve label is positioned at the start or end of the curve
AUTO
always selects GREEDY.
GREEDY
specifies the Greedy method for managing label collision. The Greedy method
tries different placement combinations in order to find an optimal approximation
that avoids collisions. Label placement using this method is often less optimal
than label placement using the Simulated Annealing (SA) method. However,
depending on the number of data points and the potential for label collisions, the
Greedy process can be significantly faster.
SA
specifies the Simulated Annealing method for managing label collision. The SA
method attempts to determine the global minimization-of-cost function, which is
based on a simulated annealing algorithm. The resulting label placement is
usually better than placement using the Greedy method. However, depending on
the number of data points and the potential for label collisions, the SA method
can be significantly slower.

Restriction For BANDPLOT and LINECHART, the SA method has no effect

on the curve labels when the CURVELABELPOSITION= option
specifies START or END.
BEGINGRAPH Statement 31

Default The value specified by the ODS GRAPHICS statement

LABELPLACEMENT= option, which is AUTO by default.

Restriction The data label placement algorithm is not aware of bar labels, curve
labels, box plot outlier labels, and marker characters. Collisions
between these elements and data labels might occur regardless of the
LABELPLACEMENT= setting.

Interactions This option overrides the ODS GRAPHICS statement

LABELPLACEMENT= option.

This option affects a plot’s labels only when

DATALABELPOSITION=AUTO is in effect for that plot.

The data label font size might be reduced in order to avoid

overlapping labels and markers. Starting with the third maintenance
release of SAS 9.4, when a broken axis is used, the data-label font size
is not scaled during label placement.

OPAQUE=TRUE | FALSE
specifies whether the graph background is opaque or transparent.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
When this option is set to FALSE, the graph background is transparent.

Default TRUE

Restriction A transparent background is supported only by the PNG, EMF, PDF,

and SVG output formats. The PS output format supports a transparent
background when the graph is rendered as a PNG image. It does not
support a transparent background when the graph is rendered as vector-
graphics output.

Interaction When this option is set to FALSE, the BACKGROUNDCOLOR=

option has no effect.

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the graph border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.
32 Chapter 2 • BEGINGRAPH Statement

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default Padding for all sides is 10 pixels.

Note The default units for dimension are pixels.

See “dimension” on page 1340

SAPLACEMENTOPTS=(placement-options)
specifies the options for the label-placement algorithm when
LABELPLACEMENT=SA. Placement options can be any of the following:
MAXITERATIONS=positive-integer
specifies the maximum number of iterations for the SA label-placement
algorithm.

Default 100

WEIGHTS=(keyword-number-list)
specifies the relative weight to give to a particular cost when determining the best
label position. The keyword number list is a space-separated list of keyword =
number pairs.
The following keywords can be used:

LABEL assigns a weight to the overlapping of labels

MARKER assigns a weight to the overlapping of markers and
labels
OUTOFBOUND assigns a weight to labels that are out-of-bounds or
clipped
PRIORITY assigns a weight to the priority of each potential label
position
OBSTACLE assigns a weight to the overlapping of labels with drop
lines, needles, reference lines, series lines, step lines,
and vector lines
The higher the number, the more weight is assigned to the specified cost. For
example, if MARKER is given more weight than OBSTACLE, avoiding marker
collisions is given a higher priority than avoiding line collisions.

Default A weight of 1.0 is assigned to each keyword

Example saplacementopts=(maxiterations=100
weights=(LABEL=2.0 OBSTACLE=10.0))

SEED=positive-integer
specifies a random number seed for the Simulated Annealing algorithm.

Default 1234567

Range 0–2147483646 (2 31–1), where 0 specifies the current Java time as the
seed value
BEGINGRAPH Statement 33

Restriction This option applies only when LABELPLACEMENT=SA.

SUBPIXEL=AUTO | OFF | ON
specifies whether subpixel rendering is used for drawing smooth curved lines or for
spacing bars more precisely.
Note: Starting with the third maintenance release of SAS 9.4, this option controls
subpixel rendering only for image output. For vector-graphics output, subpixel
rendering is always enabled.
AUTO
the system determines whether to use subpixel rendering. In the second
maintenance release of SAS 9.4 and in earlier releases, the system uses the
default rendering for the rendering technology. Starting with the third
maintenance release of SAS 9.4, for image output, if the SUBPIXEL= option is
explicitly set in an ODS GRAPHICS statement, the system honors its setting.
Otherwise, the system determines whether to use subpixel rendering based on the
following criteria:
• If a SCATTERPLOT or SCATTERPLOTMATRIX statement is used,
subpixel rendering is OFF for the graph.
• If neither a SCATTERPLOT nor a SCATTERPLOTMATRIX statement is
used, subpixel rendering is turned ON for the graph if one or more of the
following statements is also used:

BANDPLOT DENSITYPLOT LOESSPLOT

BARCHART HEATMAP PARETOLINE
BARCHARTPARM HEATMAPPARM PBSPLINEPLOT
BOXPLOT HIGHLOWPLOT POLYGONPLOT
BOXPLOTPARM HISTOGRAM REGRESSIONPLOT
BUBBLEPLOT HISTOGRAMPARM SERIESPLOT
CONTOURPLOTPARM LINECHART WATERFALLCHART
• For all other cases, subpixel rendering is turned OFF.
OFF
never uses subpixel rendering.

Note OFF is valid starting with the third maintenance release of SAS 9.4.

ON
always uses subpixel rendering, when applicable, for image output when
rendering graphs.

Default AUTO

Restrictions In the second maintenance release of SAS 9.4 and in earlier releases,
subpixel rendering can be used only for the following statements:
BANDPLOT, BARCHART, BARCHARTPARM, DENSITYPLOT,
LINECHART, LOESSPLOT, PBSPLINEPLOT,
REGRESSIONPLOT, and SERIESPLOT. Starting with the third
maintenance release of SAS 9.4, subpixel rendering can be used for
all plots and charts.

Starting with the third maintenance release of SAS 9.4, this option is
ignored for vector-graphics output.
34 Chapter 2 • BEGINGRAPH Statement

Requirement Anti-aliasing must be enabled for this option to have any effect.

Interaction Starting with the third maintenance release of SAS 9.4, this option
overrides the SUBPIXEL= option in the ODS GRAPHICS statement.

Tips If anti-aliasing is disabled, use the ANTIALIAS=ON option in the

ODS GRAPHICS statement to enable it.

Anti-aliasing is disabled automatically when the resources required

for anti-aliasing exceed a preset threshold. When anti-aliasing is
disabled for all or part of a graph, subpixel rendering is disabled for
the entire graph. A note is written to the SAS log that provides
information about how to use the ANTIALIASMAX= option in an
ODS GRAPHICS statement to re-enable anti-aliasing.

See “Using Subpixel Rendering” in SAS Graph Template Language:

User's Guide

“ODS GRAPHICS Statement” in SAS ODS Graphics: Procedures

Guide for information about the ANTIALIAS= and
ANTIALIASMAX= options.

Details

About the BEGINGRAPH Statement

All template definitions in the Graphics Template Language must start with a
BEGINGRAPH statement and end with an ENDGRAPH statement. Within a
BEGINGRAPH block, one and only one GTL layout block is required. It can be a
LATTICE, GRIDDED, OVERLAY, OVERLAYEQUATED, OVERLAY3D, REGION,
DATALATTICE, or DATAPANEL layout block. This layout block and its nested
layouts, if any, must contain at least one plot statement. It can contain other nested
layout blocks.
The GTL global statements apply to the entire template and can include ENTRYTITLE
and ENTRYFOOTNOTE statements, attribute maps, draw statements, conditional
statements, and so on. Any of these global statements can precede or follow the GTL
layout block.

Changing the Size of Your Graph

By default, graphs are rendered at 640px by 480px (4:3 aspect ratio). To change the
output size for a single graph, use the DESIGNWIDTH= and DESIGNHEIGHT=
options in the BEGINGRAPH statement for that graph. For example, the template in the
“Example Program” on page 36 uses DESIGNHEIGHT= to change the graph height to
320px. To prevent the graph width from automatically scaling to preserve the 4:3 aspect
ratio, it uses DESIGNWIDTH= to maintain the 640px width. In this instance, the setting
renders each graph cell as a 320px by 320px square. (The cells are square in this case,
but the resulting cell size depends on the graph definition and would not be the same for
all graphs.)
Note: To change the graph sizes for all templates in the current SAS session, you can
use the WIDTH= and HEIGHT= options in the ODS GRAPHICS statement. Size
settings in the ODS GRAPHICS statement override size settings in the
BEGINGRAPH statement and remain in effect unless they are changed in another
ODS GRAPHICS statement. You can also use WIDTH= and HEIGHT= settings in
the graph style to modify the graphs sizes across template definitions. Be aware,
BEGINGRAPH Statement 35

however, that if you explicitly manage the graph output size, then the graph elements
might be scaled so that the size specification is honored.
The following template defines a square graph (equal height and width, 1:1 aspect ratio)
by setting the design width equal to the internal default height (480px). The setting is
made with DESIGNWIDTH=DEFAULTDESIGNHEIGHT:
Note: A “square graph” means that the output graph’s width and height are equal. That
does not imply that the X and Y axis lengths are equal if the graph contains only one
cell.
proc template;
define statgraph squareplot;
dynamic title xvar yvar;
begingraph / designwidth=defaultDesignHeight;
entrytitle title;
layout overlayequated / equatetype=square;
scatterplot x=xvar y=yvar;
regressionplot x=xvar y=yvar;
endlayout;
endgraph;
end;
run;

If this template is executed with the following GRENDER procedure statement, then a
480px by 480px graph is created:

proc sgrender data=mydata template="squareplot" ;

dynamic title="Square Plot" xvar="time1" yvar="time2";
run;

If the ODS GRAPHICS statement’s WIDTH= or HEIGHT= options change the render
width or render height, then the squareplot template’s 1:1 aspect ratio would still be
honored. Thus, both of the following GRENDER procedure statements would create a
550px by 550px graph:
ods graphics / width=550px;
proc sgrender data=mydata template="squareplot" ;
dynamic title="Square Plot" xvar="time1" yvar="time2";
run;

ods graphics / height=550px;

proc sgrender data=mydata template="squareplot" ;
dynamic title="Square Plot" xvar="time1" yvar="time2";
run;
36 Chapter 2 • BEGINGRAPH Statement

Example: BEGINGRAPH Statement

The following graph was generated by the “Example Program” on page 36:

Example Program
The BEGINGRAPH statement block is a required outermost container for any graph
template. One of its purposes is to support options that apply to the entire graph. For
example, the default graph size that a template produces is typically 640x480 pixels. If
you need a different size, then you can declare the alternative size on this statement. To
do so, use the DESIGNWIDTH= option, or the DESIGNHEIGHT= option, or both. This
program shows one way to set the width and height of two graph cells to be equal.
proc template;
define statgraph begingraph;
dynamic XVAR YVAR;
begingraph / designwidth=640px designheight=320px;
layout lattice / columns=2;
layout overlayequated / equatetype=square;
entry "Linear Regression Fit" /
valign=top texttattrs=(weight=bold);
scatterplot x=XVAR y=YVAR / datatransparency=0.5;
regressionplot x=XVAR y=YVAR;
endlayout;
layout overlayequated / equatetype=square;
entry "Loess Fit" /
valign=top texttattrs=(weight=bold);
scatterplot x=XVAR y=YVAR / datatransparency=0.5;
loessplot x=XVAR y=YVAR;
endlayout;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.iris template=begingraph;

Example: BEGINGRAPH Statement 37

dynamic title="Square Plot"

xvar="SepalLength" yvar="SepalWidth";
run;
38 Chapter 2 • BEGINGRAPH Statement
39

Part 3

Layout Statements

Chapter 3
Summary of Layout Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Chapter 4
Layout Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
40
41

Chapter 3
Summary of Layout Statements

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Single-cell Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Multi-cell Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Data-driven Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Legend Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Overview
Layout blocks always begin with the LAYOUT keyword followed by a keyword
indicating the purpose of the layout. All layout blocks end with an ENDLAYOUT
statement.
The following sections summarize the available layouts. To learn more about a layout,
see the chapter devoted to that layout.
42 Chapter 3 • Summary of Layout Statements

Single-cell Layouts

Graphics Allowed
Layout and Cells
(Description) Produced Comments Example

OVERLAY (Single 2-D (1 cell) General purpose

Cell) layout for
superimposing 2-D
plots

OVERLAYEQUATE 2-D (1 cell) Specialized

D (Single Cell) OVERLAY with
equated axes

PROTOTYPE 2-D (1 cell) Specialized

(Single Cell) LAYOUT used only
as child layout of
DATAPANEL or
DATALATTICE

REGION (Single 2-D (1 cell) General purpose

Cell) layout for displaying
a single-cell plot that
does not use axes

OVERLAY3D 3-D (1 cell) General purpose 3-D

(Single Cell) layout for
superimposing 3-D
plots.
Data-driven Layouts 43

Multi-cell Layouts

Graphics Allowed
Layout and Cells
(Description) Produced Comments Example

LATTICE (Advanced 2-D (1 or more cells) All cells must be

Multi-cell) predefined. Axes can
be shared across
columns or rows and
be external to grid.
Many grid labeling
and alignment
features.

GRIDDED (Simple 2-D (1 or more cells) All cells must be

Multi-cell) predefined. Axes
independent for each
cell. Very simple
multi-cell container.

Data-driven Layouts

Graphics Allowed
Layout and Cells
(Description) Produced Comments Example

DATAPANEL 2-D (1 or more cells) Displays a panel of

(Classification Panel) similar graphs based
on data subsets by
classification
variable(s). Number
of cells is based on
crossings of n
classification
variable(s).

DATALATTICE 2-D (1 or more cells) Displays a panel of

(Classification Panel) similar graphs based
on data subsets by
classification
variable(s). Number
of cells is based on
crossings of 1 or 2
classification
variables.
44 Chapter 3 • Summary of Layout Statements

Legend Layout

Layout Cells Produced Comments

GLOBALLEGEND 1 cell for a legend Specialized layout for creating a compound

legend that contains multiple discrete legends.
45

Chapter 4
Layout Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
LAYOUT DATALATTICE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
LAYOUT DATAPANEL Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
LAYOUT GLOBALLEGEND Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
LAYOUT GRIDDED Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
LAYOUT LATTICE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
LAYOUT OVERLAY Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
LAYOUT OVERLAYEQUATED Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
LAYOUT OVERLAY3D Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
LAYOUT PROTOTYPE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
LAYOUT REGION Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
INNERMARGIN Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

Dictionary

LAYOUT DATALATTICE Statement

Creates a grid of graphs based on one or two classification variables and a graphical prototype. By default,
a separate instance of the prototype (a data cell) is created for each possible combination of the
classification variables.
Restriction: You can specify only one LAYOUT PROTOTYPE block in the LAYOUT
DATALATTICE block. If you specify more than one, then only the last prototype block
specified is honored. The remaining prototype blocks are ignored.
Requirement: You must specify at least one ROWVAR= option or one COLUMNVAR= option. You
can specify both.
46 Chapter 4 • Layout Statements

Syntax
LAYOUT DATALATTICE ROWVAR=class-variable
COLUMNVAR=class-variable </option(s)>;
LAYOUT PROTOTYPE </options>;
GTL-statements;
ENDLAYOUT;
<SIDEBAR </options>;
GTL-statements;
ENDSIDEBAR;>
ENDLAYOUT;
LAYOUT DATALATTICE COLUMNVAR=class-variable </option(s)>;
layout-prototype-block ;
<sidebar-block(s)> ;
ENDLAYOUT;
LAYOUT DATALATTICE ROWVAR=class-variable </option(s)>;
layout-prototype-block ;
<sidebar-block(s)> ;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CELLHEIGHTMIN=dimension
specifies the minimum height of a cell in the grid.
CELLWIDTHMIN=dimension
specifies the minimum width of a cell in the grid.
COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.
COLUMNHEADERS=TOP | BOTTOM | BOTH
specifies where to position the outside column heading.
HEADERBACKGROUNDCOLOR=style-reference | color
specifies the background color of the cell headers.
HEADERBORDER=TRUE | FALSE
specifies whether a border is drawn around the header cells.
HEADERLABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the data labels.
HEADERLABELLOCATION=OUTSIDE | INSIDE
indicates whether the cell header is placed within each cell (INSIDE) or as
row and column headers external to the lattice (OUTSIDE).
HEADEROPAQUE=TRUE | FALSE
specifies whether the background for cell headers is opaque (TRUE) or
transparent (FALSE).
LAYOUT DATALATTICE Statement 47

HEADERPACK=TRUE | FALSE
specifies whether the header cells are consolidated into a comma-separated
list in order to save space.
HEADERSEPARATOR="string"
specifies one or more characters to place between each value in the cell
header when HEADERPACK=TRUE.
HEADERSPLITCOUNT=positive-integer
specifies the number of headers to consolidate on a header line before
splitting the text to the next line.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled, depending on the nesting
levels of the layouts that are used.
SORTORDER=(role-sort-list)
specifies the order of the cells along the columns and rows.

Axis options
COLUMN2AXISOPTS=(axis-options)
specifies X2-axis options for all columns.
COLUMN2DATARANGE=AUTO | UNIONALL | UNION
specifies how the X2-axes of instances of the graph-prototype are scaled.
COLUMNAXISOPTS=(axis-options)
specifies X-axis options for all columns.
COLUMNDATARANGE=AUTO | UNIONALL | UNION
specifies how the X-axes of instances of the graph-prototype are scaled.
ROW2AXISOPTS=(axis-options)
specifies Y2-axis options for all rows.
ROW2DATARANGE=AUTO | UNIONALL | UNION
specifies how the Y2-axes of instances of the graph-prototype are scaled.
ROWAXISOPTS=(axis-options)
specifies Y-axis options for all rows.
ROWDATARANGE=AUTO | UNIONALL | UNION
specifies how the Y-axes of instances of the graph-prototype are scaled.

Inset options
INSET=(variable-list)
specifies what information is displayed in an inset.
INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information.

Layout options
COLUMNS=integer
specifies the number of columns in the layout.
COLUMNWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the columns widths.
48 Chapter 4 • Layout Statements

HEADERLABELDISPLAY=NAMEVALUE | VALUE | NONE

specifies the content of the cell headers.
INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether to include grid cells for crossings of the ROWVAR and
COLUMNVAR variables that contain a missing value.
PANELNUMBER=positive-integer
specifies the number of the panel to produce.
ROWGUTTER=dimension
specifies the amount of empty space between the rows.
ROWHEADERS=RIGHT | LEFT | BOTH
specifies where to position the outside row heading.
ROWS=integer
specifies the number of rows in the layout.
ROWWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the row heights.
SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the empty cells in a partially filled
grid.
START=TOPLEFT | BOTTOMLEFT
indicates whether to start populating the grid from the top left or bottom left
corner.

Required Arguments
You must specify at least one of the following arguments. You can specify both.
ROWVAR=class-variable
specifies the classification variable for the rows. One row of cells is created for each
unique value of the row class variable.

See ROWS= option and “Managing Rows and Columns” on page 67

COLUMNVAR=class-variable
specifies the classification variable for the columns. One column is created of each
unique value of the column class variable.

See COLUMNS= option and “Managing Rows and Columns” on page 67

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect for the color to be seen. By default,

OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
LAYOUT DATALATTICE Statement 49

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

CELLHEIGHTMIN=dimension
specifies the minimum height of a cell in the grid. Use this option in conjunction
with the CELLWIDTHMIN= option to set the minimum cell size.
The overall size of the panel is constrained by the HEIGHT= and WIDTH= options
in the ODS GRAPHICS statement. As the number of cells in the grid increases, the
size of each cell decreases. At some point the cell becomes so small that a
meaningful graph cannot be rendered. This option sets the minimum height threshold
for all cells. If the actual cell height becomes smaller, then no panel is drawn.

Default 100px

See “dimension” on page 1340

CELLWIDTHMIN=dimension
specifies the minimum width of a cell in the grid. Use this option in conjunction with
the CELLHEIGHTMIN= option to set the minimum cell size.
The overall size of the panel is constrained by the HEIGHT= and WIDTH= options
in the ODS GRAPHICS statement. As the number of cells in the grid increases, the
size of each cell decreases. At some point the cell becomes so small that a
meaningful graph cannot be rendered. This option sets the minimum width threshold
for all cells. If the actual cell width becomes smaller, then no panel is drawn.

Default 100px

See “dimension” on page 1340

COLUMNAXISOPTS=(axis-options)
specifies X-axis options for all columns.

Restriction Axis options must be enclosed in parentheses and separated by spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” for a list of options..

COLUMN2AXISOPTS=(axis-options)
specifies X2-axis options for all columns.
50 Chapter 4 • Layout Statements

Restriction This option is needed only if you use a plot statement that supports a
secondary X2 axis. If you do not use that statement’s XAXIS= option
to map data to the X2 axis, then this option is ignored. For more
information about how data are mapped to the axes, see “Plot Data
Are Mapped to a Designated Axis” on page 876

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” for a list of options.

COLUMNDATARANGE=AUTO | UNIONALL | UNION

specifies how the X-axes of instances of the graph-prototype are scaled.
AUTO
selects the X-axis scale based on the COLUMNWEIGHT= option and the
column axis type, as follows:
• When COLUMNWEIGHT=EQUAL (default), UNIONALL is selected.
• When COLUMNWEIGHT=PROPORTIONAL and the column axis is
discrete, UNION is selected. Otherwise, UNIONALL is selected.
UNIONALL
scales the X-axis data ranges across all layout columns and panels (when
PANELNUMBER= is in effect).
UNION
scales the X-axis data ranges separately for each column on a per-panel basis.
The scaling does not span multiple panels.

Default AUTO

Tip Use the COLUMNAXISOPTS= option to control shared axis features.

See The COLUMNWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

COLUMN2DATARANGE=AUTO | UNIONALL | UNION

specifies how the X2-axes of instances of the graph-prototype are scaled.
AUTO
automatically selects the X2-axis scale based on the COLUMNWEIGHT= option
and the column axis type, as follows:
• When COLUMNWEIGHT=EQUAL (default), UNIONALL is selected.
• When COLUMNWEIGHT=PROPORTIONAL and the column axis is
discrete, UNION is selected. Otherwise, UNIONALL is selected.
UNIONALL
scales the X2-axis data ranges across all layout columns and panels (when
PANELNUMBER= is in effect).
UNION
scales the X2-axis data ranges separately for each column on a per-panel basis.
The scaling does not span multiple panels.
LAYOUT DATALATTICE Statement 51

Default AUTO

Interaction This option is needed only if you use a plot statement that supports a
secondary X2 axis. If you do not use that statement’s XAXIS= option
to map data to the X2 axis, then this option is ignored. For more
information about how data are mapped to the axes, see “Plot Data Are
Mapped to a Designated Axis” on page 876

Tip Use the COLUMN2AXISOPTS= option to control shared axis features.

See The COLUMNWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.

Default 0

Note If there are n columns, then there are n-1 gutters.

See “dimension” on page 1340

COLUMNHEADERS=TOP | BOTTOM | BOTH

specifies where to position the outside column heading.
TOP
specifies that column heading text appears at the top of the layout.
BOTTOM
specifies that column heading text appears at the bottom of the layout.
BOTH
specifies that column heading text alternates between the top and bottom of the
layout column by column.

Default TOP

Interaction HEADERLABELLOCATION= OUTSIDE must be set for this

option to have any effect.

COLUMNS=integer
specifies the number of columns in the layout.

Defaults If this option is not specified, then the number of columns is

dynamically adjusted to equal the number of classifier values for the
COLUMVAR= variable.

If this option is specified, that many columns are created. If the

number of COLUMNVAR classifier values is greater than the
specified number of columns, then no graph is created for some
classifier values. If the number of classifier values is smaller than the
specified number of columns, then extra empty columns are created.

Interactions The overall grid size is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the grid size grows, the
52 Chapter 4 • Layout Statements

cell size shrinks. To control the minimum size of a cell use the
CELLHEIGHTMIN= and CELLWIDTHMIN= options.

The START= option affects the how the columns are populated.

The PANELNUMBER= option enables you to create multiple smaller

grids that completely partition the classifier values.

COLUMNWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the columns widths.
EQUAL
all columns have equal width.
PROPORTIONAL
each column has a width that is proportional to the number of discrete midpoint
values that it contains.

Restriction At least one column axis must be discrete in order for

PROPORTIONAL to have any effect. Otherwise, EQUAL is used.

Interactions When COLUMNDATARANGE=UNIONALL, PROPORTIONAL

is ignored and EQUAL is used.

When PROPORTIONAL is in effect,

COLUMNDATARANGE=AUTO is interpreted as UNION.

If all of the following conditions are true, then the discrete axis is
used to proportion the columns: PROPORTIONAL is in effect,
both the X and X2 axes are used, and only one of the two axes is
discrete. If both axes are discrete, then the X axis is used to
proportion the columns.

When PROPORTIONAL is in effect, the OFFSETMIN= and

OFFSETMAX= axis options are ignored in
COLUMNAXISOPTS= and COLUMN2AXISOPTS=. The axis
offsets are always set to one-half of the midpoint spacing
regardless of plot type.

Default EQUAL

HEADERBACKGROUNDCOLOR=style-reference | color
specifies the background color of the cell headers.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphHeaderBackground:Color style reference.

Interaction HEADEROPAQUE= TRUE must be in effect for the color to be seen.

HEADERBORDER=TRUE | FALSE
specifies whether a border is drawn around the header cells.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
LAYOUT DATALATTICE Statement 53

Default TRUE

Tip The border attributes are controlled by the GraphBorderLines style

element.

See “boolean ” for other Boolean values that you can use.

HEADERLABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

HEADERLABELDISPLAY=NAMEVALUE | VALUE | NONE

specifies the content of the cell headers.
NAMEVALUE
displays the classification variable name and value as a name=value pair in each
cell header.

Example If the classification variables are Country and Product, then

HEADERLABEL=NAMEVALUE produces cell headers such as the
following:

Country=CANADA
Product=TABLE

VALUE
displays the classification variable value only in each cell header.

Example If the classification variables are Country and Product, then

HEADERLABEL=VALUE produces cell headers such as the
following:

CANADA
TABLE

NONE
suppresses the cell headers.

Default NAMEVALUE

HEADERLABELLOCATION=OUTSIDE | INSIDE
indicates whether the cell header is placed within each cell (INSIDE) or as row and
column headers external to the lattice (OUTSIDE).

Default OUTSIDE

HEADEROPAQUE=TRUE | FALSE
specifies whether the background for cell headers is opaque (TRUE) or transparent
(FALSE).

Default TRUE
54 Chapter 4 • Layout Statements

Interaction When this option is set to FALSE, the background color for cell headers
is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

HEADERPACK=TRUE | FALSE
specifies whether the header cells are consolidated into a comma-separated list in
order to save space.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
The following figure shows the effect of HEADERPACK= on the cell headers in one
row of a data lattice. The data lattice contains classification variables Country and Year.

Default FALSE

Note If the length of the cell header text exceeds the available width, then the
text is truncated.

Tip If truncation occurs, then use the HEADERSPLITCOUNT= option to split

the cell header text into multiple lines.

See “boolean ” on page 1339 for other Boolean values that you can use.

HEADERSEPARATOR="string"
specifies one or more characters to place between each value in the cell header when
HEADERPACK=TRUE.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.

Default A comma followed by a space

Interaction This option is ignored when HEADERPACK=FALSE.

HEADERSPLITCOUNT=positive-integer
specifies the number of headers to consolidate on a header line before splitting the
text to the next line.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
LAYOUT DATALATTICE Statement 55

Default The cell header text is not split

Interaction This option is ignored when HEADERPACK=FALSE.

Note If the length of the cell header text exceeds the available width, then the
text is truncated.

Tip Use the HEADERSEPARATOR= option to specify a different

separator.

INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether to include grid cells for crossings of the ROWVAR and
COLUMNVAR variables that contain a missing value.
TRUE
any crossing of the class variables that includes a missing value produces a row
or column of cells in the grid.
FALSE
any crossing of the class variables that includes a missing value does not produce
a row or columns of cells in the grid.
By default, missing class values are included in the classification. When the data
contains missing classification values, cells are created for the missing classes. The
classification headers for the missing values are blank for missing string values or a
dot for missing numeric values. You can use the
INCLUDEMISSINGCLASS=FALSE option to exclude the missing values. If you
want to keep the missing values, then you can create a format that specifies more
meaningful headings for the missing classes. For an example, see “Missing Class
Values” in SAS Graph Template Language: User's Guide.
Note: ODS Graphics does not support Unicode values in user-defined formats in the
second maintenance release of SAS 9.4 and in earlier releases. Starting with the
third maintenance release of SAS 9.4, ODS Graphics supports Unicode values in
user-defined formats only if they are preceded by the (*ESC*) escape sequence.
Example: "(*ESC*){unicode beta}". ODS Graphics does not support an
escape character that is defined in an ODS ESCAPECHAR statement in user-
defined formats.

Default TRUE

See “boolean ” on page 1339 for other Boolean values that you can use.

INSET=(variable-list)
specifies what information is displayed in an inset. The variable-list defines one or
more variables whose names and values appear as a small table in the data cells. The
variables can be either numeric or character. Variable names are separated by spaces.

Requirement No predefined information is available for the inset. You must create
the desired inset information as part of your input data. See “Creating
Your Inset Data” on page 68.

Note The variable values are associated with the data cells by data order.
That is, the first observation from all the variables in variable-list are
used in the first data cell, the second observation from all variables in
variable-list are used in the second data cell, and so on. If a value is
missing for an observation, then the corresponding name-value pair is
skipped in the affected data cell.
56 Chapter 4 • Layout Statements

Tip The location and appearance of the inset is controlled by the

INSETOPTS= option.

See “Adding Insets to Your Graph” in SAS Graph Template Language:

User's Guide

INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information. The appearance
options can be any one or more of the following values:
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the inset is automatically aligned within the layout.
NONE
does not automatically align the inset. This inset’s position is set by the
HALIGN= and VALIGN=appearance-options.
AUTO
attempts to center the inset in the area that is farthest from any surrounding
markers. Data cells might have different inset placements.
(location-list)
restricts the inset’s possible locations to those locations in the specified
location-list, and uses the location-list position that least collides with the
data cell’s other graphics features. The location-list is a space-separated list
that can contain any of these locations: TOPLEFT, TOP, TOPRIGHT, LEFT,
CENTER, RIGHT, BOTTOMLEFT, BOTTOM, and BOTTOMRIGHT.

Example AUTOALIGN=(TOPRIGHT TOPLEFT)

Default NONE

Interaction When AUTOALIGN=AUTO or (location-list), the HALIGN= and

VALIGN= options are ignored.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the inset background.
style-reference
specifies a style reference in the form style-element : style-attribute. Only the
COLOR and CONTRASTCOLOR style attributes are valid.

Default The background is transparent. No color is assigned.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the inset.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

CONTENTDISPLAY=LABELVALUE | VALUE
specifies whether the variable information that is displayed in the inset includes
the column label and value, or only the column value.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.
If a label is not assigned to a column, then the column name is used as the label
for that column. Consider the following inset data:
LAYOUT DATALATTICE Statement 57

F
Obs _TYPE_ Value Pr > F
1 SS1 94.359 <.0001

The following figure shows the effect that the CONTENTDISPLAY= option has
on the content of an inset that displays this data.

Default LABELVALUE

Tip Use the SEPARATOR= option to specify a separator other than the
default blank space.

DATASCHEME=LIST | MATCHED
specifies the scheme that was used to merge the inset information into the
analysis data.
LIST
one-to-one merging (no BY statement) was used to merge the inset and
analysis data. The variable values are associated with the cells of the data
lattice by using data order. That is, the inset variable values in the first
observation are used in the inset for the first cell, the inset variable values in
the second observation are used in the inset for the second cell, and so on.
MATCHED
match-merging (using a BY statement) was used to merge the inset and
analysis data.

Default LIST

Tip MATCHED is the preferred data scheme for merging the inset and
analysis data.

See “Adding Insets to Classification Panels” in SAS Graph Template

Language: User's Guide

HALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment of the inset.

Default LEFT

Interaction This option has an effect only when this layout is nested within a
region layout or when this layout is nested in an overlay-type layout
and AUTOALIGN=NONE.

OPAQUE=TRUE | FALSE
specifies whether the inset background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When OPAQUE=FALSE, the background color is not used.

58 Chapter 4 • Layout Statements

See “boolean ” on page 1339 for other Boolean values that you can use.

SEPARATOR="string"
specifies a new separator for the column label and value.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Default A blank space

Interaction This option is ignored when CONTENTDISPLAY=VALUE.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the entire inset, excluding the title.

Default The GraphDataText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TITLE="string"
specifies a title for the inset. The title is added at the top of the inset and spans
the full inset width.

Note Space is not reserved for the title when this value is not specified.

Tip Text properties for the title string can be specified with TITLEATTRS=.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the inset’s title string.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

VALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment of the inset.

Default TOP

Interaction This option has effect only when this layout is nested within a
region layout or when this layout is nested in an overlay-type layout
and AUTOALIGN=NONE.

Requirements The options must be enclosed in parentheses.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent (FALSE).

Default FALSE
LAYOUT DATALATTICE Statement 59

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0
60 Chapter 4 • Layout Statements

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

PANELNUMBER=positive-integer
specifies the number of the panel to produce. This option enables you to partition a
large grid into a number of smaller grids under these conditions:
• You set a grid size explicitly (ROWS= and COLUMNS= when ROWVAR and
COLUMNVAR variables are specified; ROWS= when only a ROWVAR variable
is specified; COLUMNS= when only a COLUMNVAR variable is specified)
• The grid size is smaller in one or both of the dimensions of the default
dynamically generated grid.
• You execute the template N times and increment the panel number each time. N
is determined by CEIL(all rows * all columns / grid rows * grid columns).

Default 1

Example Suppose ROWVAR=R (R has 10 unique values) and COLUMNVAR=C

(C has 11 unique values). The dynamic grid has 10 rows and 11 columns
and you would have to make the HEIGHT=and WIDTH= quite large to
enable 110 plots to be displayed. By setting some smaller grid size, say
ROWS=3 and COLUMNS=4, and by making the value of
PANELNUMBER= a dynamic or macro variable, you can create 10
panels (9 with 12 data cells and 1 with 2 data cells) that collectively
display all 110 possible crossings. You simply invoke PROC
SGRENDER or a DATA step 10 times, incrementing the dynamic value
for PANELNUMBER each time.

ROWAXISOPTS=(axis-options)
specifies Y-axis options for all rows.

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” on page 1032 for a list of options..

ROW2AXISOPTS=(axis-options)
specifies Y2-axis options for all rows.
LAYOUT DATALATTICE Statement 61

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

Interaction This option is needed only if you use a plot statement that supports a
secondary Y2 axis. If you do not use that statement’s YAXIS= option
to map data to the Y2 axis, then this option is ignored. For more
information about how data are mapped to the axes, see “Plot Data
Are Mapped to a Designated Axis” on page 876

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” on page 1032 for a list of options..

ROWDATARANGE=AUTO | UNIONALL | UNION

specifies how the Y-axes of instances of the graph-prototype are scaled.
AUTO
automatically selects the Y-axis scale based on the ROWWEIGHT= option and
the column axis type, as follows:
• When ROWWEIGHT=EQUAL (default), UNIONALL is selected.
• When ROWWEIGHT=PROPORTIONAL and the row axis is discrete,
UNION is selected. Otherwise, UNIONALL is selected.
UNIONALL
scales the Y-axis data ranges across all layout rows and panels (when
PANELNUMBER= is in effect).
UNION
scales the Y-axis data ranges separately for each row in the layout on a per-panel
basis. The scaling does not span multiple panels.

Default AUTO

Tip Use the ROWAXISOPTS= option to control shared axis features.

See The ROWWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

ROW2DATARANGE=AUTO | UNIONALL | UNION

specifies how the Y2-axes of instances of the graph-prototype are scaled.
AUTO
automatically selects the Y2-axis scale based on the ROWWEIGHT= option and
the column axis type, as follows:
• When ROWWEIGHT=EQUAL (default), UNIONALL is selected.
• When ROWWEIGHT=PROPORTIONAL and the row axis is discrete,
UNION is selected. Otherwise, UNIONALL is selected.
UNIONALL
scales the Y2-axis data ranges across all layout rows and panels (when
PANELNUMBER= is in effect).
UNION
scales the Y2-axis data ranges separately for each row in the layout on a per-
panel basis. The scaling does not span multiple panels.
62 Chapter 4 • Layout Statements

Default AUTO

Interaction This option is needed only if you use a plot statement that supports a
secondary Y2 axis. If you do not use that statement’s YAXIS= option to
map data to the Y2 axis, then this option is ignored. For more
information about how data are mapped to the axes, see “Plot Data Are
Mapped to a Designated Axis” on page 876

Tip Use the ROW2AXISOPTS= option to control shared axis features.

See The ROWWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Default 0

Note If there are n rows, then there are n-1 gutters.

See “dimension” on page 1340

ROWHEADERS=RIGHT | LEFT | BOTH

specifies where to position the outside row heading.
RIGHT
specifies that row heading appears at the right of the layout.
LEFT
specifies that row heading appears at the left of the layout.
BOTH
specifies that row heading alternates between the right and left of the layout row
by row.

Default RIGHT

Requirement HEADERLABELLOCATION= OUTSIDE must be set for this

option to have any effect.

ROWS=integer
specifies the number of rows in the layout.

Defaults If this option is not specified, then the number of rows is dynamically
adjusted to equal the number of classifier values for the ROWVAR=
variable.

If this option is specified, then the specified number of rows is created.

If the number of ROWVAR classifier values is greater than the
specified number of rows, then no graph is created for some classifier
values. If the number of classifier values is smaller than the specified
number of rows, then extra empty rows are created.

Interactions The overall grid size is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the grid size grows, the
LAYOUT DATALATTICE Statement 63

cell size shrinks. To control the minimum size of a cell use the
CELLHEIGHTMIN= and CELLWIDTHMIN= options.

The START= option affects how the rows are populated.

Tip The PANELNUMBER= option enables you to create multiple smaller

grids that completely partition the classifier values.

ROWWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the row heights.
EQUAL
all rows have equal height.
PROPORTIONAL
each row has a height that is proportional to the number of discrete midpoint
values that it contains.

Restriction At least one row axis must be discrete in order for

PROPORTIONAL to have any effect. Otherwise, EQUAL is used.

Interactions When ROWDATARANGE=UNIONALL, PROPORTIONAL is

ignored and EQUAL is used.

When PROPORTIONAL is in effect, ROWDATARANGE=AUTO

is interpreted as UNION.

If all of the following conditions are true, then the discrete axis is
used to proportion the rows: PROPORTIONAL is in effect, both
the Y and Y2 axes are used, and only one of the two axes is
discrete. When both axes are discrete, the Y axis is used to
proportion the rows.

When PROPORTIONAL is in effect, the OFFSETMIN= and

OFFSETMAX= axis options are ignored in ROWAXISOPTS= and
ROW2AXISOPTS=. The axis offsets are always set to one-half of
the midpoint spacing regardless of plot type.

Default EQUAL

SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled, depending on the nesting levels of
the layouts that are used.

Default FALSE

Note Fonts maintain their size regardless of the specifications in the nested
layouts.

See “boolean ” on page 1339 for other Boolean values that you can use.

SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the empty cells in a partially filled grid.
TRUE
skips empty cells and "snaps" the external axes to the nearest data cell, both
vertically and horizontally. Though the empty cells are not displayed, the data
cells in the grid are not enlarged to fill the area.
64 Chapter 4 • Layout Statements

FALSE
displays external axes at their normal locations, even if there are empty cells at
one or more of the locations.
Whenever the number of unique COLUMNVAR= classifier values (data cells) is not
evenly divisible by the COLUMNS= value, or the number of unique ROWVAR=
classifier values (data cells) is not evenly divisible by the ROWS= value, then one or
more panels is partially filled with data cells and padded with empty cells to
complete the grid.
Here is an example of a data lattice that consists of 4 column-data cells and 3 row-
data cells arranged in a 4-column, 2-row grid. The following figure shows the default
appearance of the last panel:

When SKIPEMPTYCELLS=TRUE, the empty padding cells of all panels are

removed and external axis ticks and tick values snap to the data cells:

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

SORTORDER=(role-sort-list)
specifies the order of the cells along the columns and rows. The role sort list is a list
of rolename=sort-order-keyword pairs, enclosed in parentheses.
rolename
a role name, which must be one of the following:

COLUMNVAR the column role.

ROWVAR the row variable role.
LAYOUT DATALATTICE Statement 65

sort-order-keyword
a sort-order keyword, which must be one of the following:

AUTO sorts using DATA for character data and

ascending unformatted for numeric data.
DATA retains the data order.
ASCENDINGFORMATTED sorts in ascending order, using the
formatted values.
DESCENDINGFORMATTED sorts in descending order, using the
formatted values.

Default AUTO for all roles.

Tip The placement of the cells within the layout also depends on the starting
location, which is controlled by the START= option.

START=TOPLEFT | BOTTOMLEFT
indicates whether to start populating the grid from the top left or bottom left corner.
If ROWVAR=R has values in the sort order 1 and 2 and COLUMNVAR=C has
values in the sort order a and b, then START=BOTTOMLEFT is populated as
follows:

START=TOPLEFT is populated as follows:

Default TOPLEFT

SIDEBAR Statement Options

ALIGN=BOTTOM | TOP | LEFT | RIGHT
specifies the sidebar’s location within the layout. You can specify up to four
SIDEBAR blocks in a LAYOUT DATALATTICE, one for each of the bottom, top,
left, and right sidebar positions.
• The LAYOUT DATALATTICE automatically aligns a sidebar with the layout
columns or rows.
• Only one statement (such as ENTRY or DISCRETELEGEND) or one layout
block (such as LAYOUT GRIDDED) is allowed in a SIDEBAR block. To create
multi-line text in a sidebar, nest ENTRY statements within a LAYOUT
GRIDDED block.

Default BOTTOM
66 Chapter 4 • Layout Statements

SPACEFILL=TRUE | FALSE
specifies whether to fill all the area of the sidebar with its contents.

Default TRUE

Tip To prevent a layout block within the sidebar from expanding to the sidebar
boundaries, set this option to FALSE.

See “boolean ” on page 1339 for other Boolean values that you can use.

Details

Statement Description
The LAYOUT DATALATTICE statement makes it easy to create a grid of graphs, based
on the values of one or two classifications variables. To create a grid that is based on
more than two classification variables, or to have more control over the grid layout, use
LAYOUT DATAPANEL instead.
By default, the number of cells in the layout is determined by the number of value
pairings that are possible for the classification values plus any empty cells needed to
complete the last row or column of the grid. The contents of each data cell are based on a
graph prototype that you specify in the graph-prototype-block. You can enhance the
display using one or more sidebar-statement-blocks. For classification variables that
have many values, you can use the COLUMNS= and or ROWS= options and the
PANELNUMBER= option to generate multiple panel displays.
Classification variables for the layout are specified on the ROWVAR= argument (to
specify a row variable), or the COLUMNVAR= argument (to specify a column), or both
arguments to specify both a column and a row variable. The graph prototype for each
data cell’s contents is specified within a “LAYOUT PROTOTYPE Block” on page 68
block, and sidebars are specified within SIDEBAR blocks. The LAYOUT PROTOTYPE
and SIDEBAR blocks are nested within the LAYOUT DATALATTICE block.
By default, the first data cell to be filled is in the layout’s top left corner. Use the
START= option to change the starting data cell to the bottom left corner.
Rather than display the header labels outside the grid, you can set
HEADERLABELLOCATION= INSIDE to display them inside the grid, as shown in the
following figure:
LAYOUT DATALATTICE Statement 67

Note: The DATALATTICE layout is designed to be the outermost layout in the

template.

Managing Rows and Columns

If you do not explicitly manage columns and rows using the COLUMNS= and ROWS=
options, then the default layout behavior is as follows:
• If both ROWVAR= and COLUMNVAR= are specified, then a data cell is created for
each of the value pairings that are possible for the classification values of the
specified variables. If the ROWVAR variable has R distinct values and the
COLUMNVAR variable has C distinct values, then the dimension of grid produced
is R x C.
• If only the ROWVAR variable is used, then an R x 1 grid is produced.
• If only the COLUMNVAR variable is used, then a 1 x C grid is produced.
If the class variable is of type character, then its values are returned in data order. To
control the ordering of the values, you can sort the input data by the classification
variables. If the class variable is of type numeric, then the values are displayed in ordinal
order.
Formats can be assigned to class variables to create classification levels (for example, an
AGEGROUPFMT. format for numeric AGE). In this case, the classification is
performed after the format is applied. For numeric data, the order is ordinal, based on the
first value in each class.
Use the INCLUDEMISSINGCLASS option to control whether cells are displayed when
any value crossing contains a missing value.
The output size does not grow automatically as the number of cells increases. To set a
panel size for the current template, use the DESIGNHEIGHT= and DESIGNWIDTH=
options in the BEGINGRAPH statement. To set a panel size for all templates in the
current SAS session, use the HEIGHT= and WIDTH= options in the ODS GRAPHICS
statement. Size settings in the ODS GRAPHICS statement override size settings in the
68 Chapter 4 • Layout Statements

BEGINGRAPH statement. The default output width is 640px, and the default output
height is 480px.
As the number of cells in the grid increases, the size of each cell decreases. At some
point the cells might become so small that a meaningful graph cannot be rendered. The
CELLHEIGHTMIN= and CELLWIDTHMIN= options set a threshold for the smallest
cell. If the actual cell height or width becomes smaller, then no panel is drawn. The
default minimum cell size is CELLHEIGHTMIN=100px and
CELLWIDTHMIN=100px.
Using the default panel size and cell size, the DATALATTICE layout accommodates a
grid of about 24 cells (6 columns by 4 rows). If you know that the number of cells is
larger, then you should increase the overall panel size, or decrease the minimum cell
size, or both. You can also use ROWS= , COLUMNS= , and PANELNUMBER= options
to partition your data so that a number of smaller grids are produced that cumulatively
show all of the value crossings.

Creating Your Inset Data

When you use the INSET= option to insert an inset, no predefined information is
available for the inset. You must create the desired inset information as part of your input
data. This is most typically done as follows:
• Create a separate data set for the inset columns making sure that the column names
are different from the other columns used in graph. The number observations of inset
data should match the number of cells in the classification panel. The ordering of the
inset observations should be the same as the population order of the classification
panel’s cells, taking into account the ROWVAR= and COLUMNVAR= arguments
and the START= option. Typically, the number of observations for the inset data is
smaller than the other input data for the graph.
• Merge the inset data set with the data set for the graph using a DATA or PROC SQL
step. Do not match-merge the observations of the two data sets (no BY processing).
The resulting data set typically has the inset columns padded with missing values.
• Use the merged data set to produce the graph, specifying the inset column names in
this option’s variable-list.

LAYOUT PROTOTYPE Block

You must specify a single LAYOUT PROTOTYPE block within the LAYOUT
DATALATTICE block, using the following syntax:
LAYOUT PROTOTYPE </option(s)>;
GTL-statement(s);
ENDLAYOUT;
The LAYOUT PROTOTYPE block determines the graphical content of each data cell
and is repeated within each data cell, based on the subsets of the classification variables.
For more information about the LAYOUT PROTOTYPE block and the list of available
options, see “LAYOUT PROTOTYPE Statement” on page 159.

SIDEBAR Blocks
A LAYOUT DATALATTICE enables you to display one or more sidebars outside of the
axes. A sidebar spans across columns or rows and is useful for displaying information
that applies to all of the columns or all of the rows. For example, sidebars are useful for
displaying a legend.
A SIDEBAR statement has the following syntax:
Example: LAYOUT DATALATTICE 69

SIDEBAR </ option(s)>;

GTL-statement(s);
ENDSIDEBAR;
The following example shows a SIDEBAR block that displays a legend at the top of the
layout grid.

sidebar / align=top;
discretelegend 'p' 'a' / across=2;
endsidebar;

Example: LAYOUT DATALATTICE

This example shows the result of using row and column classification variables. In this
case, a four-column, three-row data lattice is created:
• The classification values are placed as row or column labels by default.
• The ROWDATARANGE=UNION option assures that an axis range is computed
separately for each row using the data ranges of all Y= columns in that row. This
facilitates the visual comparison of the data cells.
• A SIDEBAR block is used to place the legend at the bottom of the lattice.
The following graph was generated by the “Example Program” on page 69:

Example Program
proc template;
70 Chapter 4 • Layout Statements

define statgraph layoutdatalattice;

begingraph;
entrytitle "Annual Furniture Sales Comparisons";
layout datalattice rowvar=country columnvar=year /
rowdatarange=union
headerlabeldisplay=value
headerbackgroundcolor=GraphAltBlock:color
rowaxisopts=(display=(tickvalues) griddisplay=on
linearopts=(tickvalueformat=dollar12.))
columnaxisopts=(display=(tickvalues)
timeopts=(tickvalueformat=monname3.));
layout prototype / cycleattrs=true;
seriesplot x=month y=TotalActual / name="Actual";
seriesplot x=month y=TotalPredict / name="Predict";
endlayout;
sidebar / align=bottom;
discretelegend "Actual" "Predict" / border=false;
endsidebar;
endlayout;
endgraph;
end;
run;

proc summary data=sashelp.prdsal2 nway;

class country year month;
var actual predict;
output out=prdsal2 sum=TotalActual TotalPredict;
run;

proc sgrender data=prdsal2 template=layoutdatalattice;

run;

LAYOUT DATAPANEL Statement

Creates a grid of graphs based on one or more classification variables and a graphical prototype. By
default, a separate instance of the prototype (a data cell) is created for each actual combination of the
classification variables.
Restriction: You can specify only one LAYOUT PROTOTYPE block in the LAYOUT DATAPANEL
block. If you specify more than one, then only the last prototype block specified is
honored. The remaining prototype blocks are ignored.
Tip: The DATAPANEL layout should be the outermost layout in the template.
LAYOUT DATAPANEL Statement 71

Syntax
LAYOUT DATAPANEL CLASSVARS=(class-var1…class-varN) </option(s)> ;
LAYOUT PROTOTYPE </option(s)>;
GTL-statements;
ENDLAYOUT;
<SIDEBAR </ option(s) >;
GTL-statement(s);
ENDSIDEBAR;
<… more-sidebar-statement-blocks …> >
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CELLHEIGHTMIN=dimension
specifies the minimum height of a cell in the grid.
CELLWIDTHMIN=dimension
specifies the minimum width of a cell in the grid.
COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.
HEADERBACKGROUNDCOLOR=style-reference | color
specifies the background color of the cell headers.
HEADERBORDER=TRUE | FALSE
specifies whether a border is drawn around the header cells.
HEADERLABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the data labels.
HEADEROPAQUE=TRUE | FALSE
specifies whether the background for cell headers is opaque (TRUE) or
transparent (FALSE).
HEADERPACK=TRUE | FALSE
specifies whether the header cells are consolidated into a comma-separated
list in order to save space.
HEADERSEPARATOR="string"
specifies one or more characters to place between each value in the cell
header when HEADERPACK=TRUE.
HEADERSPLITCOUNT=positive-integer
specifies the number of headers to consolidate on a header line before
splitting the text to the next line.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
72 Chapter 4 • Layout Statements

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled, depending on the nesting
levels of the layouts that are used.
SORTORDER=(role-sort-list)
specifies the order of the cells along the columns and rows.

Axis options
COLUMN2AXISOPTS=(axis-options)
specifies X2-axis options for all columns.
COLUMN2DATARANGE=AUTO | UNIONALL | UNION
specifies how the X2-axes of instances of the graph-prototype are scaled.
COLUMNAXISOPTS=(axis-options)
specifies X-axis options for all columns.
COLUMNDATARANGE=AUTO | UNIONALL | UNION
specifies how the X-axes of instances of the graph-prototype are scaled.
ROW2AXISOPTS=(axis-options)
specifies Y2-axis options for all rows.
ROW2DATARANGE=AUTO | UNIONALL | UNION
specifies how the Y2-axes of instances of the graph-prototype are scaled.
ROWAXISOPTS=(axis-options)
specifies Y-axis options for all rows.
ROWDATARANGE=AUTO | UNIONALL | UNION
specifies how the Y-axes of instances of the graph-prototype are scaled.

Inset options
INSET=(variable-list)
specifies what information is displayed in an inset.
INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information.

Layout options
COLUMNS=integer
specifies the number of columns in the layout.
COLUMNWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the columns widths.
HEADERLABELDISPLAY=NAMEVALUE | VALUE | NONE
specifies the content of the cell headers.
INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether to include grid cells for crossings of the CLASSVARS
variables that contain a missing value.
ROWGUTTER=dimension
specifies the amount of empty space between the rows.
ROWS=integer
specifies the number of rows in the layout.
ROWWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the row heights.

Panel options
ORDER=ROWMAJOR | COLUMNMAJOR
LAYOUT DATAPANEL Statement 73

specifies whether data cells are populated by column priority or by row

priority.
PANELNUMBER=positive-integer
specifies the number of the panel to produce.
SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the empty cells in a partially filled
grid.
SPARSE=TRUE | FALSE
specifies whether crossings of the class variables include only the crossings
in the data or all possible crossings.
START=TOPLEFT | BOTTOMLEFT
indicates whether to start populating the grid cells from the top left or bottom
left corner.

Role options
ROLENAME=(role-name-list)
specifies user-defined roles for information contained in data columns.

Required Argument
CLASSVARS=(column-list)
specifies a list of classification variables. By default, a data cell is created for each
crossing of these variables in the input data. The total number of grid cells created is
the result of a crosstabulation table of all the classification variables plus any empty
cells needed to complete the last row or column of the grid. You can request that data
cells be generated for all possible crossings, even when the class variables have no
values at those crossings. For more information, see the SPARSE= option.
If the class variable is of type character, then its values are returned in data order. To
control the ordering of the values, you can sort the input data by the classification
variables. If the class variable is of type numeric, then the values are displayed in
ordinal order.
Formats can be assigned to class variables to create classification levels (for
example, an AGEGROUPFMT. format for numeric AGE). In this case, the
classification is performed after the format is applied. For numeric data, the order is
ordinal, based on the first value in each class.
Use the INCLUDEMISSINGCLASS option to control whether cells are displayed
when any value crossing contains a missing value.
The output size does not grow automatically as the number of cells increases. To set
a panel size for the current template, use the DESIGNHEIGHT= and
DESIGNWIDTH= options in the BEGINGRAPH statement. To set a panel size for
all templates in the current SAS session, use the HEIGHT= and WIDTH= options in
the ODS GRAPHICS statement. Size settings in the ODS GRAPHICS statement
override size settings in the BEGINGRAPH statement. The default output width is
640px, and the default output height is 480px.
As the number of cells in the grid increases, the size of each cell decreases. At some
point the cells might become so small that a meaningful graph cannot be rendered.
The CELLHEIGHTMIN= and CELLWIDTHMIN= options set a threshold for the
smallest cell. If the actual cell height or width becomes smaller, then no panel is
drawn. The default minimum cell size is CELLHEIGHTMIN=100px and
CELLWIDTHMIN=100px.
Using the default panel size and cell size, the DATAPANEL layout accommodates a
grid of about 24 cells (6 columns by 4 rows). If you know that the number of cells is
74 Chapter 4 • Layout Statements

larger, then you should increase the overall panel size, or decrease the minimum cell
size, or both. You can also use ROWS= , COLUMNS= , and PANELNUMBER=
options to partition your data so that a number of smaller grids are produced that
cumulatively show all of the value crossings.

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect for the color to be seen. By default,

OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

HEADERBORDER=TRUE | FALSE
specifies whether a border is drawn around the header cells.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default TRUE

Tip The border attributes are controlled by the GraphBorderLines style

element.

See “boolean ” for other Boolean values that you can use.

CELLHEIGHTMIN=dimension
specifies the minimum height of a cell in the grid. Use this option in conjunction
with the CELLWIDTHMIN= option to set the minimum cell size.
LAYOUT DATAPANEL Statement 75

The overall size of the panel is constrained by the HEIGHT= and WIDTH= options
in the ODS GRAPHICS statement. As the number of cells in the grid increases, the
size of each cell decreases. At some point the cell becomes so small that a
meaningful graph cannot be rendered. This option sets the minimum height threshold
for all cells. If the actual cell height becomes smaller, then no panel is drawn.

Default 100px

See “dimension” on page 1340

CELLWIDTHMIN=dimension
specifies the minimum width of a cell in the grid. Use this option in conjunction with
the CELLHEIGHTMIN= option to set the minimum cell size.
The overall size of the panel is constrained by the HEIGHT= and WIDTH= options
in the ODS GRAPHICS statement. As the number of cells in the grid increases, the
size of each cell decreases. At some point the cell becomes so small that a
meaningful graph cannot be rendered. This option sets the minimum width threshold
for all cells. If the actual cell width becomes smaller, then no panel is drawn.

Default 100px

See “dimension” on page 1340

COLUMNAXISOPTS=(axis-options)
specifies X-axis options for all columns.

Restriction Axis options must be enclosed in parentheses and separated by spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” for a list of options..

COLUMN2AXISOPTS=(axis-options)
specifies X2-axis options for all columns.

Restriction This option is needed only if you use a plot statement that supports a
secondary X2 axis. If you do not use that statement’s XAXIS= option
to map data to the X2 axis, then this option is ignored. For more
information about how data are mapped to the axes, see “Plot Data
Are Mapped to a Designated Axis” on page 876

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” for a list of options.

COLUMNDATARANGE=AUTO | UNIONALL | UNION

specifies how the X-axes of instances of the graph-prototype are scaled.
AUTO
selects the X-axis scale based on the COLUMNWEIGHT= option and the
column axis type, as follows:
• When COLUMNWEIGHT=EQUAL (default), UNIONALL is selected.
• When COLUMNWEIGHT=PROPORTIONAL and the column axis is
discrete, UNION is selected. Otherwise, UNIONALL is selected.
76 Chapter 4 • Layout Statements

UNIONALL
scales the X-axis data ranges across all layout columns and panels (when
PANELNUMBER= is in effect).
UNION
scales the X-axis data ranges separately for each column on a per-panel basis.
The scaling does not span multiple panels.

Default AUTO

Tip Use the COLUMNAXISOPTS= option to control shared axis features.

See The COLUMNWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

COLUMN2DATARANGE=AUTO | UNIONALL | UNION

specifies how the X2-axes of instances of the graph-prototype are scaled.
AUTO
automatically selects the X2-axis scale based on the COLUMNWEIGHT= option
and the column axis type, as follows:
• When COLUMNWEIGHT=EQUAL (default), UNIONALL is selected.
• When COLUMNWEIGHT=PROPORTIONAL and the column axis is
discrete, UNION is selected. Otherwise, UNIONALL is selected.
UNIONALL
scales the X2-axis data ranges across all layout columns and panels (when
PANELNUMBER= is in effect).
UNION
scales the X2-axis data ranges separately for each column on a per-panel basis.
The scaling does not span multiple panels.

Default AUTO

Interaction This option is needed only if you use a plot statement that supports a
secondary X2 axis. If you do not use that statement’s XAXIS= option
to map data to the X2 axis, then this option is ignored. For more
information about how data are mapped to the axes, see “Plot Data Are
Mapped to a Designated Axis” on page 876

Tip Use the COLUMN2AXISOPTS= option to control shared axis features.

See The COLUMNWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.

Default 0

Note If there are n columns, then there are n-1 gutters.

See “dimension” on page 1340

LAYOUT DATAPANEL Statement 77

COLUMNS=integer
specifies the number of columns in the layout.

Defaults If this option is not specified and ROWS= is specified, then the
number of data cells (and columns) increases dynamically to allow all
classifier values to be presented.

If both this option and ROWS= are specified, then a grid of that size is
created, regardless of the number of classifier values. If the number of
classifier values is greater than the grid size, then no graphs are
created for some classifier values. If the number of classifier values is
small and the grid size large, then there might be empty cells created.

Interactions The overall grid size is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the grid size grows, the
cell size shrinks. To control the minimum size of a cell use the
CELLHEIGHTMIN= and CELLWIDTHMIN= options.

The START= and ORDER= options affect the how the rows are
populated.

The PANELNUMBER= option enables you to create multiple smaller

grids that completely partition the classifier values.

COLUMNWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the columns widths.
EQUAL
all columns have equal width.
PROPORTIONAL
each column has a width that is proportional to the number of discrete midpoint
values that it contains.

Restriction At least one column axis must be discrete in order for

PROPORTIONAL to have any effect. Otherwise, EQUAL is used.

Interactions When COLUMNDATARANGE=UNIONALL, PROPORTIONAL

is ignored and EQUAL is used.

When PROPORTIONAL is in effect,

COLUMNDATARANGE=AUTO is interpreted as UNION.

If all of the following conditions are true, then the discrete axis is
used to proportion the columns: PROPORTIONAL is in effect,
both the X and X2 axes are used, and only one of the two axes is
discrete. If both axes are discrete, then the X axis is used to
proportion the columns.

When PROPORTIONAL is in effect, the OFFSETMIN= and

OFFSETMAX= axis options are ignored in
COLUMNAXISOPTS= and COLUMN2AXISOPTS=. The axis
offsets are always set to one-half of the midpoint spacing
regardless of plot type.

Default EQUAL
78 Chapter 4 • Layout Statements

HEADERBACKGROUNDCOLOR=style-reference | color
specifies the background color of the cell headers.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphHeaderBackground:Color style reference.

Interaction HEADEROPAQUE= TRUE must be in effect for the color to be seen.

HEADERLABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

HEADERLABELDISPLAY=NAMEVALUE | VALUE | NONE

specifies the content of the cell headers.
NAMEVALUE
displays the classification variable name and value as a name=value pair in each
cell header.

Example If the classification variables are Country and Product, then

HEADERLABEL=NAMEVALUE produces cell headers such as the
following:

Country=CANADA
Product=TABLE

VALUE
displays the classification variable value only in each cell header.

Example If the classification variables are Country and Product, then

HEADERLABEL=VALUE produces cell headers such as the
following:

CANADA
TABLE

NONE
suppresses the cell headers.

Default NAMEVALUE

HEADEROPAQUE=TRUE | FALSE
specifies whether the background for cell headers is opaque (TRUE) or transparent
(FALSE).

Default TRUE

Interaction When this option is set to FALSE, the background color for cell headers
is not used.
LAYOUT DATAPANEL Statement 79

See “boolean ” on page 1339 for other Boolean values that you can use.

HEADERPACK=TRUE | FALSE
specifies whether the header cells are consolidated into a comma-separated list in
order to save space.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
The following figure shows the effect of HEADERPACK= on the cell headers in one
row of a data lattice. The data lattice contains classification variables Country, Year, and
Product.

Default FALSE

Note If the length of the cell header text exceeds the available width, then the
text is truncated.

Tip If truncation occurs, then use the HEADERSPLITCOUNT= option to split

the cell header text into multiple lines.

See “boolean ” on page 1339 for other Boolean values that you can use.

HEADERSEPARATOR="string"
specifies one or more characters to place between each value in the cell header when
HEADERPACK=TRUE.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.

Default A comma followed by a space

Interaction This option is ignored when HEADERPACK=FALSE.

HEADERSPLITCOUNT=positive-integer
specifies the number of headers to consolidate on a header line before splitting the
text to the next line.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
80 Chapter 4 • Layout Statements

The following figure shows how HEADERSPLITCOUNT=2 splits the cell header value
in a data panel of classification variables Country, Year, and Product.

Default The cell header text is not split

Interaction This option is ignored when HEADERPACK=FALSE.

Note If the length of the cell header text exceeds the available width, then the
text is truncated.

Tip Use the HEADERSEPARATOR= option to specify a different

separator.

INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether to include grid cells for crossings of the CLASSVARS variables
that contain a missing value.
TRUE
any crossing of the class variables that includes a missing value produces a row
or column of cells in the grid.
FALSE
any crossing of the class variables that includes a missing value does not produce
a row or columns of cells in the grid.
By default, missing class values are included in the classification. When the data
contains missing classification values, cells are created for the missing classes. The
classification headers for the missing values are blank for missing string values or a
dot for missing numeric values. You can use the
INCLUDEMISSINGCLASS=FALSE option to exclude the missing values. If you
want to keep the missing values, then you can create a format that specifies more
meaningful headings for the missing classes. For an example, see “Missing Class
Values” in SAS Graph Template Language: User's Guide.
Note: ODS Graphics does not support Unicode values in user-defined formats in the
second maintenance release of SAS 9.4 and in earlier releases. Starting with the
third maintenance release of SAS 9.4, ODS Graphics supports Unicode values in
user-defined formats only if they are preceded by the (*ESC*) escape sequence.
Example: "(*ESC*){unicode beta}". ODS Graphics does not support an
escape character that is defined in an ODS ESCAPECHAR statement in user-
defined formats.

Default TRUE

See “boolean ” on page 1339 for other Boolean values that you can use.

INSET=(variable-list)
specifies what information is displayed in an inset. The variable-list defines one or
more variables whose names and values appear as a small table in the data cells. The
variables can be either numeric or character. Variable names are separated by spaces.
LAYOUT DATAPANEL Statement 81

Requirement No predefined information is available for the inset. You must create
the desired inset information as part of your input data. See “Creating
Your Inset Data” on page 68.

Note The variable values are associated with the data cells by data order.
That is, the first observation from all the variables in variable-list are
used in the first data cell, the second observation from all variables in
variable-list are used in the second data cell, and so on. If a value is
missing for an observation, then the corresponding name-value pair is
skipped in the affected data cell.

Tip The location and appearance of the inset is controlled by the

INSETOPTS= option.

See “Adding Insets to Your Graph” in SAS Graph Template Language:

User's Guide

INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information. The appearance
options can be any one or more of the following values:
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the inset is automatically aligned within the layout.
NONE
does not automatically align the inset. This inset’s position is set by the
HALIGN= and VALIGN=appearance-options.
AUTO
attempts to center the inset in the area that is farthest from any surrounding
markers. Data cells might have different inset placements.
(location-list)
restricts the inset’s possible locations to those locations in the specified
location-list, and uses the location-list position that least collides with the
data cell’s other graphics features. The location-list is a space-separated list
that can contain any of these locations: TOPLEFT, TOP, TOPRIGHT, LEFT,
CENTER, RIGHT, BOTTOMLEFT, BOTTOM, and BOTTOMRIGHT.

Example AUTOALIGN=(TOPRIGHT TOPLEFT)

Default NONE

Interaction When AUTOALIGN=AUTO or (location-list), the HALIGN= and

VALIGN= options are ignored.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the inset background.
style-reference
specifies a style reference in the form style-element : style-attribute. Only the
COLOR and CONTRASTCOLOR style attributes are valid.

Default The background is transparent. No color is assigned.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the inset.
82 Chapter 4 • Layout Statements

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

CONTENTDISPLAY=LABELVALUE | VALUE
specifies whether the variable information that is displayed in the inset includes
the column label and value, or only the column value.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.
If a label is not assigned to a column, then the column name is used as the label
for that column. Consider the following inset data:
F
Obs _TYPE_ Value Pr > F
1 SS1 94.359 <.0001

The following figure shows the effect that the CONTENTDISPLAY= option has
on the content of an inset that displays this data.

Default LABELVALUE

Tip Use the SEPARATOR= option to specify a separator other than the
default blank space.

DATASCHEME=LIST | MATCHED
specifies the scheme that was used to merge the inset information into the
analysis data.
LIST
one-to-one merging (no BY statement) was used to merge the inset and
analysis data. The variable values are associated with the cells of the data
lattice by using data order. That is, the inset variable values in the first
observation are used in the inset for the first cell, the inset variable values in
the second observation are used in the inset for the second cell, and so on.
MATCHED
match-merging (using a BY statement) was used to merge the inset and
analysis data.

Default LIST

Tip MATCHED is the preferred data scheme for merging the inset and
analysis data.

See “Adding Insets to Classification Panels” in SAS Graph Template

Language: User's Guide

HALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment of the inset.
LAYOUT DATAPANEL Statement 83

Default LEFT

Interaction This option has an effect only when this layout is nested within a
region layout or when this layout is nested in an overlay-type layout
and AUTOALIGN=NONE.

OPAQUE=TRUE | FALSE
specifies whether the inset background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When OPAQUE=FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

SEPARATOR="string"
specifies a new separator for the column label and value.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Default A blank space

Interaction This option is ignored when CONTENTDISPLAY=VALUE.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the entire inset, excluding the title.

Default The GraphDataText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TITLE="string"
specifies a title for the inset. The title is added at the top of the inset and spans
the full inset width.

Note Space is not reserved for the title when this value is not specified.

Tip Text properties for the title string can be specified with TITLEATTRS=.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the inset’s title string.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

VALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment of the inset.

Default TOP
84 Chapter 4 • Layout Statements

Interaction This option has effect only when this layout is nested within a
region layout or when this layout is nested in an overlay-type layout
and AUTOALIGN=NONE.

Requirements The options must be enclosed in parentheses.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row priority.
ROWMAJOR
fills the data cells by rows, from the starting position.
COLUMNMAJOR
fills the data cells by columns, from the starting position.

Default ROWMAJOR

Interaction The starting point for rendering data cells is controlled by the START=
option. See the START= option for examples.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.
LAYOUT DATAPANEL Statement 85

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

PANELNUMBER=positive-integer
specifies the number of the panel to produce. This option enables you to partition a
large grid into a number of smaller sized grids under these conditions:
• You set a grid size explicitly ( ROWS= and COLUMNS= options).
• The grid size (gridrows x gridcolumns) is smaller than the total number of
classifier levels.
• You execute the template N times and increment the panel number each time. N
is determined by CEIL(total-classification-levels / gridrows x
gridcolumns).
86 Chapter 4 • Layout Statements

Default 1

Example Suppose there are two classifiers (CLASS1 has 10 unique values and
CLASS2 has 11 unique values). By setting some smaller grid size, say
ROWS=3 and COLUMNS=4, and making the value of
PANELNUMBER= a dynamic or macro variable, you can create 10
panels (9 panels with 12 data cells and 1 panel with 2 data cells) that
collectively display all 110 possible crossings. You simply invoke PROC
SGRENDER or a DATA step 10 times, incrementing the dynamic value
for PANELNUMBER each time.

ROLENAME=(role-name-list)
specifies user-defined roles for information contained in data columns. The role
name list is a space-separated list role-name=column pairs.

Default no user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles.

Example The following example assigns the column Obs to the user-defined
role TIP1.
ROLENAME=(TIP1=OBS)

ROWAXISOPTS=(axis-options)
specifies Y-axis options for all rows.

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” on page 1032 for a list of options..

ROW2AXISOPTS=(axis-options)
specifies Y2-axis options for all rows.

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

Interaction This option is needed only if you use a plot statement that supports a
secondary Y2 axis. If you do not use that statement’s YAXIS= option
to map data to the Y2 axis, then this option is ignored. For more
information about how data are mapped to the axes, see “Plot Data
Are Mapped to a Designated Axis” on page 876

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” on page 1032 for a list of options..

ROWDATARANGE=AUTO | UNIONALL | UNION

specifies how the Y-axes of instances of the graph-prototype are scaled.
AUTO
automatically selects the Y-axis scale based on the ROWWEIGHT= option and
the column axis type, as follows:
• When ROWWEIGHT=EQUAL (default), UNIONALL is selected.
• When ROWWEIGHT=PROPORTIONAL and the row axis is discrete,
UNION is selected. Otherwise, UNIONALL is selected.
LAYOUT DATAPANEL Statement 87

UNIONALL
scales the Y-axis data ranges across all layout rows and panels (when
PANELNUMBER= is in effect).
UNION
scales the Y-axis data ranges separately for each row in the layout on a per-panel
basis. The scaling does not span multiple panels.

Default AUTO

Tip Use the ROWAXISOPTS= option to control shared axis features.

See The ROWWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

ROW2DATARANGE=AUTO | UNIONALL | UNION

specifies how the Y2-axes of instances of the graph-prototype are scaled.
AUTO
automatically selects the Y2-axis scale based on the ROWWEIGHT= option and
the column axis type, as follows:
• When ROWWEIGHT=EQUAL (default), UNIONALL is selected.
• When ROWWEIGHT=PROPORTIONAL and the row axis is discrete,
UNION is selected. Otherwise, UNIONALL is selected.
UNIONALL
scales the Y2-axis data ranges across all layout rows and panels (when
PANELNUMBER= is in effect).
UNION
scales the Y2-axis data ranges separately for each row in the layout on a per-
panel basis. The scaling does not span multiple panels.

Default AUTO

Interaction This option is needed only if you use a plot statement that supports a
secondary Y2 axis. If you do not use that statement’s YAXIS= option to
map data to the Y2 axis, then this option is ignored. For more
information about how data are mapped to the axes, see “Plot Data Are
Mapped to a Designated Axis” on page 876

Tip Use the ROW2AXISOPTS= option to control shared axis features.

See The ROWWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Default 0

Note If there are n rows, then there are n-1 gutters.

See “dimension” on page 1340

88 Chapter 4 • Layout Statements

ROWS=integer
specifies the number of rows in the layout.

Defaults If this option is not specified and COLUMNS= is specified, then the
number of data cells (and rows) increases dynamically to allow all
classifier values to be presented.

If both this option and COLUMNS= are specified, then a grid of that
size is created, regardless of the number of classifier values. If the
number of classifier values is greater than the grid size, then no graphs
are created for some classifier values. If the number of classifier
values is small and the grid size large, then there might be empty cells
created.

Interactions The overall grid size is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the grid size grows, the
cell size shrinks. To control the minimum size of a cell use the
CELLHEIGHTMIN= and CELLWIDTHMIN= options.

The START= and ORDER= options affect how the rows are
populated.

Tip The PANELNUMBER= option enables you to create multiple smaller

grids that completely partition the classifier values.

ROWWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the row heights.
EQUAL
all rows have equal height.
PROPORTIONAL
each row has a height that is proportional to the number of discrete midpoint
values that it contains.

Restriction At least one row axis must be discrete in order for

PROPORTIONAL to have any effect. Otherwise, EQUAL is used.

Interactions When ROWDATARANGE=UNIONALL, PROPORTIONAL is

ignored and EQUAL is used.

When PROPORTIONAL is in effect, ROWDATARANGE=AUTO

is interpreted as UNION.

If all of the following conditions are true, then the discrete axis is
used to proportion the rows: PROPORTIONAL is in effect, both
the Y and Y2 axes are used, and only one of the two axes is
discrete. When both axes are discrete, the Y axis is used to
proportion the rows.

When PROPORTIONAL is in effect, the OFFSETMIN= and

OFFSETMAX= axis options are ignored in ROWAXISOPTS= and
ROW2AXISOPTS=. The axis offsets are always set to one-half of
the midpoint spacing regardless of plot type.

Default EQUAL
LAYOUT DATAPANEL Statement 89

SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled, depending on the nesting levels of
the layouts that are used.

Default FALSE

Note Fonts maintain their size regardless of the specifications in the nested
layouts.

See “boolean ” on page 1339 for other Boolean values that you can use.

SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the empty cells in a partially filled grid.
TRUE
skips empty cells and "snaps" the external axes to the nearest data cell, both
vertically and horizontally. Though the empty cells are not displayed, the data
cells in the grid are not enlarged to fill the area.
FALSE
displays external axes at their normal locations, even if there are empty cells at
one or more of the locations.
Whenever the total number of classifier crossings (data cells) is not evenly divisible
by the panel size (columns * rows), the last panel is partially filled with data cells
and padded with empty cells to complete the grid.
Here is an example of a data panel that consists of 16 data cells arranged in a 4-
column, 3-row grid. The following figure shows the default appearance of the last
panel:

When SKIPEMPTYCELLS=TRUE, the empty padding cells of the last panel are
removed and external axis ticks and tick values snap to the data cells:
90 Chapter 4 • Layout Statements

Note that SKIPEMPTYCELLS=TRUE removes only the empty padding cells on the
last panel. It does not remove any data cells that have no crossing values and
therefore no graph (these data cells are displayed when SPARSE=TRUE).

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

SORTORDER=(role-sort-list)
specifies the order of the cells along the columns and rows. The role sort list is a list
of rolename=sort-order-keyword pairs, enclosed in parentheses.
rolename
a role that is defined by the ROLENAME= option.

Note Roles that are not class variables are ignored.

sort-order-keyword
a sort-order keyword, which must be one of the following:

AUTO sorts using DATA for character data and

ascending unformatted for numeric data.
DATA retains the data order.
ASCENDINGFORMATTED sorts in ascending order, using the
formatted values.
DESCENDINGFORMATTED sorts in descending order, using the
formatted values.

Default AUTO for all roles.

Tip The placement of the cells within the layout also depends on the starting
location, which is controlled by the START= option.

SPARSE=TRUE | FALSE
specifies whether crossings of the class variables include only the crossings in the
data or all possible crossings.
LAYOUT DATAPANEL Statement 91

FALSE
specifies that data cells are created only for crossings of the class variables that
are in the data.
TRUE
specifies that the number of data cells is the product of the unique values for each
classification variable.
By default, if a crossing of the class variables has a missing value as part of the data,
then a data cell is created for it.
Here is an example of a classification panel where the classification variables are
COUNTRY and STATE. There are 3 distinct values of COUNTRY (Canada, Mexico,
and U.S.A.) Within Canada and Mexico there are 4 states, and within U.S.A. there
are 8 states. All state names are unique to each country. Therefore, there are 16
unique STATE values and 48 unique crossings of COUNTRY and STATE, but there
are data for only 16 of the crossings.
Assume that a data panel layout is created with COLUMNS=6 and SPARSE=TRUE,
meaning to display all possible crossings. This is what the first row would look like.
Blank data cells are added whenever there are no data values for a crossing:

When SPARSE=FALSE the crossings of the classifiers with no data are

automatically removed. This compacts the display:

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

START=TOPLEFT | BOTTOMLEFT
indicates whether to start populating the grid cells from the top left or bottom left
corner. If ORDER=ROWMAJOR (the default) and START=TOPLEFT (the default),
then a 2 row 2 column grid is populated as shown in the following figure.
92 Chapter 4 • Layout Statements

If ORDER=ROWMAJOR (the default) and START=BOTTOMLEFT, then a 2 row 2

column grid is populated as shown in the following figure.

If ORDER=COLUMNMAJOR and START=BOTTOMLEFT, then a 2 row 2 column

grid is populated as shown in the following figure.

If ORDER=COLUMNMAJOR and START=TOPLEFT, then a 2 row 2 column grid

is populated as shown in the following figure.

Default TOPLEFT

SIDEBAR Optional Argument

ALIGN=BOTTOM | TOP | LEFT | RIGHT
specifies the sidebar’s location within the layout. You can specify up to four
SIDEBAR blocks in a LAYOUT DATAPANEL, one for each of the bottom, top, left,
and right sidebar positions.
• The LAYOUT DATAPANEL automatically aligns a sidebar with the layout
columns or rows.
• Only one statement (such as ENTRY or DISCRETELEGEND) or one layout
block (such as LAYOUT GRIDDED) is allowed in a SIDEBAR block. To create
multi-line text in a sidebar, nest ENTRY statements within a LAYOUT
GRIDDED block.
LAYOUT DATAPANEL Statement 93

Default BOTTOM

SPACEFILL=TRUE | FALSE
specifies whether to fill all the area of the sidebar with its contents.

Default TRUE

Tip To prevent a layout block within the sidebar from expanding to the sidebar
boundaries, set this option to FALSE.

See “boolean ” on page 1339 for other Boolean values that you can use.

Details

Statement Description
The LAYOUT DATAPANEL statement creates a grid of graphs, based on the values of
one or more classifications variables. The main differences between this layout and the
DATALATTICE layout is that this layout supports more than two classification
variables, and it provides more control over the grid layout.
By default, the number of cells in the layout is determined by a crosstabulation table of
all the classification variables plus any empty cells needed to complete the last row or
column of the grid. The contents of each data cell are based on a graph prototype that
you specify in the graph-prototype-block. You can enhance the display using one or
more sidebar-statement-blocks. For classification variables that have many values, you
can use the COLUMNS= option or the ROWS= option, or both with the
PANELNUMBER= option to generate multiple panel displays.
The order of the value pairings for the classification variables is determined by the order
that the variables are specified on the CLASSVARS= argument. The last named
variable’s values vary most rapidly (like nested DO loops). Variable values are always
returned in data order.
By default, the first data cell to be filled is in the layout’s top left corner, and data cells
are filled from left-to-right, top-to-bottom. Use the START= option to change the
starting data cell to the bottom left corner, and use the ORDER= option to determine
whether data cells fill by column or by row. See the START= option for illustrations on
how START= and ORDER= interact to manage the fill sequence for data cells.
Note: The DATAPANEL layout is designed to be the outermost layout in the template.

Prototype Block
You must specify a single graph-prototype-block within the LAYOUT DATAPANEL
block, using the following syntax:
LAYOUT PROTOTYPE </option(s)>;
GTL-statements;
ENDLAYOUT;
The graph-prototype-block determines the graphical content of each data cell and is
repeated within each data cell, based on the subsets of the classification variables.
For more information about the LAYOUT PROTOTYPE block and the list of available
options, see “LAYOUT PROTOTYPE Statement” on page 159.
94 Chapter 4 • Layout Statements

Sidebar Blocks
A LAYOUT DATAPANEL enables you to display sidebars outside of the axis areas. A
sidebar spans across columns or rows and is useful for displaying information that
applies to all of the columns or all of the rows. For example, sidebars are useful for
displaying a legend.
A SIDEBAR statement has the following syntax:
SIDEBAR </ option(s) >;
GTL-statement(s);
ENDSIDEBAR;
The following example shows a SIDEBAR block that displays a legend at the top of the
layout grid.

sidebar / align=top;
discretelegend 'p' 'a' / across=2;
endsidebar;

Creating Your Inset Data

When you use the INSET= option to insert an inset, no predefined information is
available for the inset. You must create the desired inset information as part of your input
data. This is most typically done as follows (see the chapter on classification panels and
the chapter on insets in SAS Graph Template Language: User's Guide for complete
examples):
• Create a separate data set for the inset columns making sure that the column names
are different from the other columns used in graph. The number observations of inset
data should match the number of cells in the classification panel. The ordering of the
inset observations should be the same as population order of the cells of the
classification panel, taking into account the CLASSVARS= argument and the
ORDER= and START= options. Typically, the number of observations for the inset
data is smaller than the other input data for the graph.
• Merge the inset data set with the data set for the graph using a DATA or PROC SQL
step. Do not match-merge the observations of the two data sets (no BY processing).
The resulting data set typically has the inset columns padded with missing values.
• Use the merged data set to produce the graph, specifying the inset column names in
this option’s variable-list.

Example: LAYOUTDATAPANEL Statement

This example shows a four-column, three-row data panel using two classification
variables. With this layout, each data cell is subsetted and labeled with the values of the
classification variables.
• The ROWDATARANGE=UNION option assures that an axis range is computed
separately for each row using the data ranges of all Y= columns in that row. This
facilitates the visual comparison of the data cells.
• A SIDEBAR block is used to place the legend at the bottom of the lattice.
Example: LAYOUTDATAPANEL Statement 95

Example Graph
The following graph was generated by the “Example Program” on page 95:

Example Program
proc template;
define statgraph layoutdatapanel;
begingraph;
entrytitle "Annual Furniture Sales Comparisons";
layout datapanel classvars=(country year) /
columns=4 rows=3 rowdatarange=union
headerlabeldisplay=value
headerbackgroundcolor=GraphAltBlock:color
rowaxisopts=(display=(tickvalues) griddisplay=on
linearopts=(tickvalueformat=dollar12.))
columnaxisopts=(display=(tickvalues)
timeopts=(tickvalueformat=monname3.));
layout prototype / cycleattrs=true;
seriesplot x=month y=TotalActual / name="Actual";
seriesplot x=month y=TotalPredict / name="Predict";
endlayout;
sidebar / align=top;
discretelegend "Actual" "Predict" / border=false;
endsidebar;
endlayout;
endgraph;
end;
run;

proc summary data=sashelp.prdsal2 nway;

96 Chapter 4 • Layout Statements

class country year month;

var actual predict;
output out=prdsal2 sum=TotalActual TotalPredict;
run;

proc sgrender data=prdsal2 template=layoutdatapanel;

run;

LAYOUT GLOBALLEGEND Statement

Creates a compound legend containing multiple discrete legends positioned at the bottom of a graph.
Restrictions: Only one global legend is allowed in a graph.
The LAYOUT GLOBALLEGEND statement must be placed directly inside the
BEGINGRAPH block. It is not valid outside of the BEGINGRAPH block.
Continuous legends are not supported inside the global legend.
When the LAYOUT GLOBALLEGEND block is used, all of the template's legend
statements must be specified within the LAYOUT GLOBALLEGEND block. Any
legend statements that are specified outside of the LAYOUT GLOBALLEGEND block
are ignored.
See: “DISCRETELEGEND and MERGEDLEGEND Statements” on page 1109

Syntax
LAYOUT GLOBALLEGEND </option(s)>;
<discreteLegend-statement(s) | mergedLegend-statement(s)>;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
DISPLAYCLIPPED=TRUE | FALSE
specifies whether the global legend is displayed when any portion of its
nested legends cannot be fully rendered because of space constraints.
GUTTER=dimension
specifies the gap between nested layouts.
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
TYPE=ROW | COLUMN
specifies whether nested legends are arranged into a single row or column.
WEIGHTS=UNIFORM | PREFERRED | (weight-list)
specifies the preferred space allocation for the nested legends.

Legend title options

LAYOUT GLOBALLEGEND Statement 97

LEGENDTITLEPOSITION=LEFT | TOP
specifies the position of each nested legend’s title.
TITLE=“string”
specifies a title for the global legend.
TITLEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the global legend title.

Location options
HALIGN=CENTER | LEFT | RIGHT
specifies the layout’s horizontal alignment within the graph area that is
defined by the BEGINGRAPH block.

Optional Arguments
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

DISPLAYCLIPPED=TRUE | FALSE
specifies whether the global legend is displayed when any portion of its nested
legends cannot be fully rendered because of space constraints. When the graph size
is reduced, parts of a nested legend (title, legend symbol, or legend value) might be
clipped (truncated). When clipping occurs and this option is FALSE, the entire global
legend is removed from the graph and the space for it is reclaimed by the remainder
of the graph. When this option is TRUE, the global legend always appears, even if
some parts of the nested legends have been clipped.

Default FALSE

Interaction This option overrides any DISPLAYCLIPPED option that is set on its
nested legend statements.

See “boolean ” on page 1339 for other Boolean values that you can use.

GUTTER=dimension
specifies the gap between nested layouts.

Default 0
98 Chapter 4 • Layout Statements

Note The default units for dimension are pixels.

See “dimension” on page 1340

HALIGN=CENTER | LEFT | RIGHT

specifies the layout’s horizontal alignment within the graph area that is defined by
the BEGINGRAPH block.

Default CENTER

Note When CENTER is in effect and the outermost layout is an overlay-type

layout, the global legend is centered below the wall area if it can fit within
the wall width.

LEGENDTITLEPOSITION=LEFT | TOP
specifies the position of each nested legend’s title. Specifying LEFT places each title
to the left of the legend items for that legend. Specifying TOP places each title above
the legend items for that legend.

Default LEFT

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Notes The default units for dimension are pixels.

Starting with the first maintenance release of SAS 9.4, the default padding
between the global legend and the plot area (including the axes) is
increased to 10 pixels. If the new default padding is not desirable, then use
this option to adjust it.
LAYOUT GLOBALLEGEND Statement 99

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

TITLE=“string”
specifies a title for the global legend.

Default No title is displayed for the global legend.

Restriction The string must be enclosed in quotation marks.

Tip The title for the global legend is independent of the titles for its nested
legends.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the global legend title.

Default The GraphLabelText style element.

Interaction For this option to have any effect, the TITLE= option must also be
specified.
100 Chapter 4 • Layout Statements

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TYPE=ROW | COLUMN
specifies whether nested legends are arranged into a single row or column.

Default ROW

Interaction When this option is set to ROW, the relative width of each legend is
determined by the setting for the WEIGHTS= option.

WEIGHTS=UNIFORM | PREFERRED | (weight-list)

specifies the preferred space allocation for the nested legends.
UNIFORM
allocates an equal amount of space for all nested legends.
PREFERRED
allocates the preferred amount of space for each nested legend.
(weight-list)
a space-separated list of preferred space allocations, enclosed in parentheses. The
list can combine numbers with the keyword PREFERRED. Each number is a
proportional weight for the corresponding nested legend (the weights do not have
to sum to 1.0). Keyword PREFERRED specifies that the corresponding nested
legend should be allocated its preferred space. The order of the weights that are
specified in the list should correspond to the order of the legend statements that
are nested in the GLOBALLEGEND layout.

Default UNIFORM

Restriction The option is supported only for TYPE=ROW.

Tip When a weight-list is specified, all the legends using PREFERRED get
their preferred space. Any remaining space is divided among the
legends, in proportion to the numeric values specified in the weight-list.

Details
A global legend layout can contain multiple discrete or merged legends. Continuous
legends are not supported inside the global legend block.
A global legend is placed at the bottom of the graph, just above the footnote(s). All of
the discrete or merged legend statements that are nested within the global legend block
are arranged into a single row or column, depending on the setting for the TYPE=
option.
Depending on the outermost layout type and the legend content, the legend is centered
on the graph wall area or on the graph output area. For example, if the outermost layout
is an overlay layout, then when positioning the legend, the GLOBALLEGEND
statement first attempts to center the legend on the graph wall area. If that position
causes the legend to be clipped, then it then attempts to center the legend on the entire
output area instead. In that case, the legend might appear to be slightly off-center with
respect to the graph.
Only one global legend block is permitted in a graph. The block must be located within
the BEGINGRAPH block, but outside of the outermost layout block.
Example: LAYOUT GLOBALLEGEND Statement 101

When a global legend block is used, only the legend statements that are specified within
the global legend block are displayed in the graph. Any legend statements that are
specified outside of the global legend block are ignored.

Example: LAYOUT GLOBALLEGEND Statement

The following graph was generated by the “Example Program” on page 101:

Example Program
proc template;
define statgraph globallegend;
begingraph;
entrytitle "Prediction Ellipses";
layout overlay;
scatterplot x=petallength y=petalwidth / group=species name="sp";
ellipse x=petallength y=petalwidth / type=predicted alpha=0.2
name="p80" legendlabel="80%" outlineattrs=graphconfidence;
ellipse x=petallength y=petalwidth / type=predicted alpha=0.05
name="p95" legendlabel="95%" outlineattrs=graphconfidence2;
endlayout;
layout globalLegend / type=column title="Sample Global Legend";
discretelegend "sp" / title="Species:";
discretelegend "p80" "p95" / title="Predictions:";
endLayout;
endgraph;
end;
run;
102 Chapter 4 • Layout Statements

proc sgrender data=sashelp.iris template=globallegend;

run;

LAYOUT GRIDDED Statement

Assembles the results of nested GTL-statements into a grid.
Note: Empty cells might be omitted from the grid. See “Details” on page 109.

Syntax
LAYOUT GRIDDED </option(s)>;
GTL-statements;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
COLUMNGUTTER=dimension
specifies the amount of empty space between the columns.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled, depending on the nesting
levels of the layouts that are used.

Grid options
COLUMNS=integer
specifies the number of columns in the layout.
ROWS=integer
specifies the number of rows in the layout.

Layout options
ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Legend options
LOCATION=INSIDE | OUTSIDE
LAYOUT GRIDDED Statement 103

specifies whether the legend appears inside or outside the plot area when
nested within an overlay-type layout.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether this layout is automatically aligned within its parent when
nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies this layout’s horizontal alignment within its parent when nested
within an overlay-type or region layout.
VALIGN=CENTER | TOP | BOTTOM | number
specifies this layout’s vertical alignment within its parent when nested within
an overlay-type or region layout.

Panel options
ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row
priority.

Optional Arguments
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether this layout is automatically aligned within its parent when nested
within an overlay-type layout.
NONE
does not automatically align the layout within its overlay-type parent layout. This
layout’s position within its parent layout is therefore set by the HALIGN= and
VALIGN= options.
AUTO
within the overlay-type parent layout, attempts to center the layout in the area
that is farthest from any surrounding data point markers. This option is available
only if the parent layout contains a scatter plot. Otherwise, it is ignored.
(location-list)
within the parent layout, restricts the layout’s possible locations to those
locations in the specified location-list, and uses the location-list position that
least collides with the parent layout’s other graphics features. The location-list is
a space-separated list that can contain any of these locations: TOPLEFT, TOP,
TOPRIGHT, LEFT, CENTER, RIGHT, BOTTOMLEFT, BOTTOM, and
BOTTOMRIGHT.

Default NONE

Restriction This option is ignored if this layout statement is the outermost layout
or if the parent layout is not an overlay-type layout.

Interactions When this option is not NONE and the parent layout is an overlay-
type layout, the HALIGN= and VALIGN= options are ignored.

This option is ignored if LOCATION= OUTSIDE.

See “LAYOUT OVERLAY Statement” on page 142 for more information

about how child positions are determined in an overlay-type layout.
104 Chapter 4 • Layout Statements

BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect for the color to be seen. By default,

OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

COLUMNGUTTER=dimension
specifies the amount of empty space between the columns.

Default 0

Note If there are n columns, then there are n-1 gutters.

See “dimension” on page 1340

COLUMNS=integer
specifies the number of columns in the layout. This option is used to create a grid
with a fixed number of columns, without concern for how many rows. For example,
the following settings ensure that columns 1 and 2 in the first row are filled with
content, as shown in the figure:

layout gridded / columns=2 order=rowmajor border=true ;

entry '1' /border=true;
entry '2' /border=true;
entry '3' /border=true;
endlayout;
LAYOUT GRIDDED Statement 105

Defaults If ORDER= ROWMAJOR, then the default is 1.

If ORDER=COLUMNMAJOR, then as many columns are created as

needed to satisfy the ROWS= request.

Restriction Assuming ORDER=ROWMAJOR, if COLUMNS=n and there are m

cells defined, and n > m, then only m columns are created (there are n -
m cells with zero size).

HALIGN=CENTER | LEFT | RIGHT | number

specifies this layout’s horizontal alignment within its parent when nested within an
overlay-type or region layout.
number
specifies the horizontal alignment as a fraction of the parent container’s width.

Range 0 to 1, where 0 is all the way to the left and 1 is all the way to the
right.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER

Restriction This option has effect only when this layout is nested within a region
layout, or when this layout is nested in an overlay-type layout and
AUTOALIGN=NONE.

See “LAYOUT OVERLAY Statement” on page 142 for more information

about how child positions are determined in an overlay-type or region
layout.

LOCATION=INSIDE | OUTSIDE
specifies whether the legend appears inside or outside the plot area when nested
within an overlay-type layout.

Default INSIDE

Restriction This option has effect only when the GRIDDED layout block appears
within an overlay-type layout. For more information about how child
positions are determined in an overlay-type layout, see “LAYOUT
OVERLAY Statement” on page 136.

Interactions If this option is set to OUTSIDE, then the HALIGN= and VALIGN=
options must specify a keyword (LEFT, RIGHT, or CENTER). The
number setting for the alignment is invalid when the layout is
positioned outside of the plot area.

The actual position is determined by this option’s setting plus the

settings for the AUTOALIGN= or HALIGN= and VALIGN= options.
106 Chapter 4 • Layout Statements

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row priority.
ROWMAJOR
fills all the columns in a row, from left to right, before going to the next row.
COLUMNMAJOR
fills all the rows in a column, from top to bottom, before going to the next
column.

Default ROWMAJOR

Requirements If this option is set to COLUMNMAJOR, then the ROWS= option

must be specified to indicate how many rows to fill before wrapping
to the next column. The default number of rows is 1.

If this option is set to ROWMAJOR, then the COLUMNS= option

must be specified to indicate how many columns to fill before
wrapping to the next column. The default number of columns is 1.

Interactions The ROWS= option is ignored when ORDER=ROWMAJOR.

The COLUMNS= option is ignored when

ORDER=COLUMNMAJOR.

See ROWS= on page 108

COLUMNS= on page 104

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
LAYOUT GRIDDED Statement 107

BOTTOM=dimension specifies the amount of extra space to add to the

bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Default 0
108 Chapter 4 • Layout Statements

Note If there are n rows, then there are n-1 gutters.

See “dimension” on page 1340

ROWS=integer
specifies the number of rows in the layout. This option is used to create a grid with a
fixed number of rows, without concern for how many columns. For example, the
following settings ensure that rows 1 and 2 in the first column are filled with content,
as shown in the figure:

layout gridded / rows=2 order=columnmajor border=true ;

entry '1' /border=true;
entry '2' /border=true;
entry '3' /border=true;
endlayout;

Defaults If ORDER= COLUMNMAJOR, then the default is 1.

If ORDER=ROWMAJOR, then this option is ignored and as many

rows are created as needed to satisfy the COLUMNS= request.

Restriction Assuming ORDER=COLUMNMAJOR, if ROWS=n and there are m

cells defined, and n > m, then only m rows are created (there are n - m
cells with zero size).

SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled, depending on the nesting levels of
the layouts that are used.

Default FALSE

Note Fonts maintain their size regardless of the specifications in the nested
layouts.

See “boolean ” on page 1339 for other Boolean values that you can use.

VALIGN=CENTER | TOP | BOTTOM | number

specifies this layout’s vertical alignment within its parent when nested within an
overlay-type or region layout.
number
specifies the vertical alignment as a fraction of the parent container’s height.

Range 0 to 1, where 0 is on the bottom and 1 is on the top.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER
Example: LAYOUT GRIDDED Statement 109

Restriction This option has effect only when this layout is nested within a region
layout, or when this layout is nested in an overlay-type layout and
AUTOALIGN=NONE.

See “LAYOUT OVERLAY Statement” on page 142 for more information

about how child positions are determined in an overlay-type or region
layout.

Details
A GRIDDED layout is commonly used to create small tables of text that are nested
within other layouts. The layout might also be used to span and center a single entry (a
legend, for example) across a set of grids. Or it might be used to display a grid of graphs
when there is no need to scale axis data ranges or align graphs across grid cells.
The GRIDDED layout automatically decides how much area to allocate to cell contents:
• Text items have a fixed size based on the amount of text and the font properties.
• Graphs take up the remaining space.
The layout’s grid size is determined by the COLUMNS= and ROWS= options. The
resulting columns and rows can be separated by areas called “gutters,” which are
controlled by the COLUMNGUTTER= and ROWGUTTER= options.
By default, the results of the GTL statements are placed into the grid sequentially from
left to right, wrapping to a new row each time the current row is filled. You can use the
ORDER= option to fill cells from top to bottom down a column. In that case, the layout
cells wrap to a new column each time the current column is filled. The GTL statements
can include text statements, plot statements, and layout blocks. Each statement or layout
block provides content for one cell in the grid.
If a statement or layout block for a grid cell does not produce any output, then the space
for that cell might not be retained as an empty cell in the grid. In that case, the empty cell
is removed, and the remaining cells (if any) fill the gap in the grid. A statement produces
no output when required data for that statement does not resolve. A layout block
produces no output when it contains no statements or when none of the statements
contained in the block produce any output.

Example: LAYOUT GRIDDED Statement

The following graph was generated by the “Example Program” on page 110:
110 Chapter 4 • Layout Statements

Example Program
The GRIDDED layout offers the best way to nest a table of information inside another
layout. In the GRIDDED layout, you can control the content, text justification, and fonts
of columns. Because this example nests the GRIDDED layout within an OVERLAY
layout, you can control where it appears within the plot area. The AUTOALIGN= option
enables you to specify a prioritized list of possible positions where the layout should be
drawn. The position actually used is the first one that avoids collision with the
histogram. Also, the GRIDDED layout is set to be opaque so that the grid lines do not
show through.
This example also illustrates a reusable template in the sense that it works for any
numeric column specified by the dynamic variable VAR. Also, SGE functions for
computing the N, MEAN, STDDEV of the column are used in the table to compute the
statistics as the template is executed.
proc template;
define statgraph inset;
dynamic VAR;
begingraph;
entrytitle "Distribution of " VAR;
layout overlay / yaxisopts=(griddisplay=on);
histogram VAR / scale=percent;
layout gridded / columns=2
autoalign=(topleft topright) border=true
opaque=true backgroundcolor=GraphWalls:color;
entry halign=left "N";
entry halign=left eval(strip(put(n(VAR),12.0)));
entry halign=left "Mean";
entry halign=left eval(strip(put(mean(VAR),12.2)));
entry halign=left "Std Dev";
entry halign=left eval(strip(put(stddev(VAR),12.2)));
endlayout;
LAYOUT LATTICE Statement 111

endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=inset;

dynamic VAR="Weight";
run;

LAYOUT LATTICE Statement

Creates a grid of graphs that automatically aligns plot areas and tick display areas across grid cells to
facilitate data comparisons among graphs.
Note: Empty cells might be omitted from the lattice. See “Cell Contents” on page 127.

Syntax
LAYOUT LATTICE </option(s)>;
GTL-statement(s) | cell-statement-block(s);
<COLUMNAXES;
COLUMNAXIS / axis-option(s);
<… more-COLUMNAXIS-statements …>
ENDCOLUMNAXES;>
<COLUMN2AXES;
COLUMNAXIS / axis-option(s);
<… more-COLUMNAXIS-statements …>
ENDCOLUMN2AXES;>
<ROWAXES;
ROWAXIS / axis-option(s);
<… more-ROWAXIS-statements …>
ENDROWAXES;>
<ROW2AXES;
ROWAXIS / axis-option(s);
<… more-ROWAXIS-statements …>
ENDROW2AXES;>
<COLUMNHEADERS;
GTL-statement(s);
ENDCOLUMNHEADERS;
<… more-header-statement-block(s) …> >
<SIDEBAR </option(s)>;
GTL-statement(s);
ENDSIDEBAR;>
<… more-sidebar-statement-blocks …> >
ENDLAYOUT;
112 Chapter 4 • Layout Statements

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled, depending on the nesting
levels of the layouts that are used.

Cell options
COLUMNWEIGHTS=UNIFORM | PREFERRED | (weight-list)
specifies the proportional width of each column relative to the overall grid
width, not including the headings and sidebars.
ROWWEIGHTS=UNIFORM | PREFERRED | (weight-list)
specifies the proportional height of each row relative to the overall grid
height, not including the headings and sidebars.

Column options
COLUMN2DATARANGE=DATA | UNION | UNIONALL
specifies how the X2-axis data ranges of graphs within the layout columns
are scaled.
COLUMNDATARANGE=DATA | UNION | UNIONALL
specifies how the X-axis data ranges of graphs within the layout columns are
scaled.

Lattice options
COLUMNS=integer
specifies the number of columns in the layout.
ROWS=integer
specifies the number of rows in the layout.
SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the unused cells in a partially filled
lattice.

Layout options
ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Location options
LAYOUT LATTICE Statement 113

AUTOALIGN=NONE | AUTO | (location-list)

specifies whether this layout is automatically aligned within its parent when
nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies this layout’s horizontal alignment within its parent when nested
within an overlay-type or region layout.
VALIGN=CENTER | TOP | BOTTOM | number
specifies this layout’s vertical alignment within its parent when nested within
an overlay-type or region layout.

Panel options
ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row
priority.

Row options
ROW2DATARANGE=DATA | UNION | UNIONALL
specifies how the Y2-axis data ranges of graphs within the layout rows are
scaled.
ROWDATARANGE=DATA | UNION | UNIONALL
specifies how the Y-axis data ranges of graphs within the layout rows are
scaled.

Optional Arguments
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether this layout is automatically aligned within its parent when nested
within an overlay-type layout.
NONE
does not automatically align the layout within its overlay-type parent layout. This
layout’s position within its parent layout is therefore set by the HALIGN= and
VALIGN= options.
AUTO
within the overlay-type parent layout, attempts to center the layout in the area
that is farthest from any surrounding data point markers. This option is available
only if the parent layout contains a scatter plot. Otherwise, it is ignored.
(location-list)
within the parent layout, restricts the layout’s possible locations to those
locations in the specified location-list, and uses the location-list position that
least collides with the parent layout’s other graphics features. The location-list is
a space-separated list that can contain any of these locations: TOPLEFT, TOP,
TOPRIGHT, LEFT, CENTER, RIGHT, BOTTOMLEFT, BOTTOM, and
BOTTOMRIGHT.

Default NONE

Restriction This option is ignored if this layout statement is the outermost layout or
if the parent layout is not an overlay-type layout.

Interaction When this option is not NONE and the parent layout is an overlay-type
layout, the HALIGN= and VALIGN= options are ignored.
114 Chapter 4 • Layout Statements

See “LAYOUT OVERLAY Statement” on page 142 for more information

about how child positions are determined in an overlay-type layout.

Example In the following example, the LATTICE layout is the child of the
OVERLAY layout. The child layout appears in either the top right or
top left position, based on which position has more “unoccupied” area.

dynamic VAR STAT1 STAT2 STAT3;

layout overlay;
histogram VAR;
layout lattice / AUTOALIGN=(TOPRIGHT TOPLEFT)
columns=1;
entry STAT1;
entry STAT2;
entry STAT3;
endlayout;
endlayout;

BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect for the color to be seen. By default,

OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

COLUMNDATARANGE=DATA | UNION | UNIONALL

specifies how the X-axis data ranges of graphs within the layout columns are scaled.
DATA
scales the X-axis data ranges separately for each cell in the layout.
LAYOUT LATTICE Statement 115

UNION
scales the X-axis data ranges separately for each column in the layout. This
setting is supported only if all plots across the column can share the same data
range and axis type. For more information, see “Plot Axis Types Must Agree on
Common Axes” on page 883.
UNIONALL
scales the X-axis data ranges across all columns in the layout. This setting is
supported only if all plots across the column can share the same data range and
axis type. For more information, see “Plot Axis Types Must Agree on Common
Axes” on page 883.

Default DATA

Interactions The data ranges are scaled only if the data values are continuous and
the graphs have the same orientation. If graphs cannot use the same
horizontal range or if COLUMNDATARANGE=DATA, then the
horizontal range of each graph is determined from the data.

If any plot statement in any cell contains a XAXIS=X2 option, then

this plot's X values are ignored whenever COLUMNDATARANGE=
is set to UNION or UNIONALL.

Tips If column axes are externalized and if a lattice cell contains a

LAYOUT OVERLAY with the XAXISOPTS= option specified, then
the XAXISOPTS option is ignored. In that case, you can use the
COLUMNAXIS statement to specify desired X-axis features. For
more information, see “Axis Statements” on page 129 .

By default, axes are always internal to the cell. To externalize column

axes, set this option to UNION or UNIONALL, and then specify a
COLUMNAXES block with as many COLUMNAXIS statements as
there are columns that contain X-axes to manage.

COLUMN2DATARANGE=DATA | UNION | UNIONALL

specifies how the X2-axis data ranges of graphs within the layout columns are
scaled.
DATA
scales the X2-axis data ranges separately for each cell in the layout.
UNION
scales the X2-axis data ranges separately for each column in the layout. This
setting is supported only if all plots across the column can share the same data
range and axis type. For more information, see “Plot Axis Types Must Agree on
Common Axes” on page 883.
UNIONALL
scales the X2-axis data ranges across all columns in the layout. This setting is
supported only if all plots across the column can share the same data range and
axis type. For more information, see “Plot Axis Types Must Agree on Common
Axes” on page 883.

Default DATA

Interactions The data ranges are scaled only if the data values are continuous and
the graphs have the same orientation. If graphs cannot use the same
116 Chapter 4 • Layout Statements

horizontal range or if COLUMN2DATARANGE=DATA, then the

horizontal range of each graph is determined from the data.

If any plot statement in any cell contains a XAXIS=X option, then this
plot's X values are ignored whenever COLUMN2DATARANGE= is
set to UNION or UNIONALL.

Tips Axes are always internal to the cell, by default. To externalize column
axes, set this option to UNION or UNIONALL, and then specify a
COLUMN2AXES block with as many COLUMNAXIS statements as
there are columns that contain X2-axes to manage.

If column axes are externalized and if a lattice cell contains a

LAYOUT OVERLAY with the X2AXISOPTS= option specified, then
the X2AXISOPTS option is ignored. In that case, you can use the
COLUMNAXIS statement to specify desired X2-axis features. For
more information, see “Axis Statements” on page 129 .

COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.

Default 0

Note If there are n columns, then there are n-1 gutters.

See “dimension” on page 1340

COLUMNS=integer
specifies the number of columns in the layout.

Defaults If ORDER= ROWMAJOR, then the default is 1.

If ORDER=COLUMNMAJOR, then as many columns are created as

are needed to satisfy the ROWS= request.

Interactions If both ROWS=n and COLUMNS=m are specified, then an n by m

grid of cells is created. If the number of statements that define cell
contents is greater than n x m, then the grid size does not expand and
some statements are not displayed. If the number of statements that
define cell contents is less than n x m, then the grid will contain empty
cells.

If this option is not defined and ORDER=COLUMNMAJOR, then the

number of columns is dynamically determined by the number of
defined cells.

COLUMNWEIGHTS=UNIFORM | PREFERRED | (weight-list)

specifies the proportional width of each column relative to the overall grid width, not
including the headings and sidebars.
UNIFORM
equally divides the total available width among all of the columns.
PREFERRED
specifies that each column gets its preferred width based on the following:
• Columns that contain one or more vertically one-dimensional plots get the
maximum preferred width from the vertically one-dimensional plots.
LAYOUT LATTICE Statement 117

• The remaining columns that do not contain vertically one-dimensional plots

get an equal amount of width from the remaining space.

Interaction If a one-dimensional box plot is specified in the preferred column,

then the box plot's BOXWIDTH= option is ignored.

Notes The PREFERRED option is used for lattice layouts that mix one-
dimensional and two-dimensional plots in the grid. It enables the
layout to compute the weights automatically for columns that
contain one-dimensional plots.

Examples of one-dimensional plots include axis tables, block plots,

fringe plots, and box plots with Y= values only.

(weight-list)
a space-separated list of column weights. The list should contain a weight for
each column, which can be expressed as one of the following:
PREFERRED
specifies that the corresponding column gets its preferred width as described
previously.

Note The PREFERRED option should be used on columns that contain

only one-dimensional plots. Using the PREFERRED keyword on
columns that contain a mix of one-dimensional and two-dimensional
plots might cause unexpected results. In that case, use a numeric
weight instead.

Tip The PREFERRED option is particularly useful for axis tables.

number
specifies that the corresponding column gets a width that is based on the
proportion of the specified number to the total of the numbers in the weight
list. For example, the following weight specifications are equivalent:
columnweights=(0.2 0.3 0.5)

columnweights=(2 3 5)

In these examples, the first column gets 20% of the available width, the
second column gets 30%, and the third column gets 50%.
If the list contains the PREFERRED keyword, then the number specifies the
proportion of the width that remains after the preferred width or widths are
calculated and subtracted from the available width.

Requirements The values in the weight list must be enclosed in parentheses.

If there are n columns in the grid, then the weight list should
contain n weights, one for each column.

All of the numbers in the list must be positive.

At least one number in the list must be nonzero.

Note The weights for all of the columns are specified as a proportion
and, as such, are not required to total 1.0.
118 Chapter 4 • Layout Statements

Example Here is an example that specifies the preferred width for the first
column, 20% of the remaining width to the second column, and
80% of the remaining width to the third column.
columnweights=(preferred 0.2 0.8)

HALIGN=CENTER | LEFT | RIGHT | number

specifies this layout’s horizontal alignment within its parent when nested within an
overlay-type or region layout.
number
specifies the horizontal alignment as a fraction of the parent container’s width.

Range 0 to 1, where 0 is all the way to the left and 1 is all the way to the
right.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER

Restriction This option has effect only when this layout is nested within a region
layout, or when this layout is nested in an overlay-type layout and
AUTOALIGN=NONE.

See “LAYOUT OVERLAY Statement” on page 142 for more information

about how child positions are determined in an overlay-type or region
layout.

Example In the following example, the LATTICE layout is the child of the
OVERLAY layout and is positioned in the OVERLAY’s top right
corner.
dynamic VAR STAT1 STAT2 STAT3;
layout overlay;
histogram VAR;
layout lattice / VALIGN=TOP HALIGN=RIGHT
columns=1;
entry STAT1;
entry STAT2;
entry STAT3;
endlayout;
endlayout;

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row priority.
ROWMAJOR
fills all the columns in a row, from left to right, before going to the next row.
LAYOUT LATTICE Statement 119

COLUMNMAJOR
fills all the rows in a column, from top to bottom, before going to the next
column.

Default ROWMAJOR

Requirements If this option is set to COLUMNMAJOR, then the ROWS= option

must be specified to indicate how many rows to fill before wrapping
to the next column. The default number of rows is 1.

If this option is set to ROWMAJOR, then the COLUMNS= option

must be specified to indicate how many columns to fill before
wrapping to the next column. The default number of columns is 1.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
120 Chapter 4 • Layout Statements

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

ROWDATARANGE=DATA | UNION | UNIONALL

specifies how the Y-axis data ranges of graphs within the layout rows are scaled.
DATA
scales the Y-axis data ranges separately for each cell in the layout.
UNION
scales the Y-axis data ranges separately for each row in the layout. This setting is
supported only if all plots down the row can share the same data range and axis
type. For more information, see “Plot Axis Types Must Agree on Common
Axes” on page 883.
UNIONALL
scales the Y-axis data ranges across all rows in the layout. This setting is
supported only if all plots down the row can share the same data range and axis
type. For more information, see “Plot Axis Types Must Agree on Common
Axes” on page 883.

Default DATA

Interactions The data ranges are scaled only if the data values are continuous and
the graphs have the same orientation. If graphs cannot use the same
vertical range or if ROWDATARANGE=DATA, then the vertical
range of each graph is determined from the data.

If any plot statement in any cell contains a YAXIS=Y2 option, then

this plot's Y values are ignored whenever ROWDATARANGE= is set
to UNION or UNIONALL.
LAYOUT LATTICE Statement 121

Tips Axes are always internal to the cell, by default. To externalize row
axes, set this option to UNION or UNIONALL, and then specify a
ROWAXES block with as many ROWAXIS statements as there are
rows that contain Y-axes to manage.

If row axes are externalized and if a lattice cell contains a LAYOUT

OVERLAY with the YAXISOPTS= option specified, then the
YAXISOPTS option is ignored. In that case, you can use the
ROWAXIS statement to specify desired Y-axis features. For more
information, see “Axis Statements” on page 129 .

ROW2DATARANGE=DATA | UNION | UNIONALL

specifies how the Y2-axis data ranges of graphs within the layout rows are scaled.
DATA
scales the Y2-axis data ranges separately for each cell in the layout.
UNION
scales the Y2-axis data ranges separately for each row in the layout. This setting
is supported only if all plots down the row can share the same data range and axis
type. For more information, see “Plot Axis Types Must Agree on Common
Axes” on page 883.
UNIONALL
scales the Y2-axis data ranges across all rows in the layout. This setting is
supported only if all plots down the row can share the same data range and axis
type. For more information, see “Plot Axis Types Must Agree on Common
Axes” on page 883.

Default DATA

Interactions The data ranges are scaled only if the data values are continuous and
the graphs have the same orientation. If graphs cannot use the same
vertical range or if ROW2DATARANGE=DATA, then the vertical
range of each graph is determined from the data.

If any plot statement in any cell contains a YAXIS=Y option, then this
plot's Y values are ignored whenever ROW2DATARANGE= is set to
UNION or UNIONALL.

Tips Axes are always internal to the cell, by default. To externalize row
axes, set this option to UNION or UNIONALL, and then specify a
ROW2AXES block with as many ROWAXIS statements as there are
rows that contain Y2-axes to manage.

If row axes are externalized and if a lattice cell contains a LAYOUT

OVERLAY with the Y2AXISOPTS= option specified, then the
Y2AXISOPTS option is ignored. In that case, you can use the
ROWAXIS statement to specify desired Y2-axis features. For more
information, see “Axis Statements” on page 129 .

ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Default 0

Note If there are n rows, then there are n-1 gutters.

122 Chapter 4 • Layout Statements

See “dimension” on page 1340

ROWS=integer
specifies the number of rows in the layout.

Defaults If ORDER=COLUMNMAJOR, then the default is 1.

If ORDER=ROWMAJOR, then as many ROWS are created as needed

to satisfy the COLUMNS= request.

Interactions If both ROWS=n and COLUMNS=m are specified, then an n by m

grid of cells is created. If the number of statements that define cell
contents is greater than n x m, then the grid size does not expand and
some statements are not displayed. If the number of statements that
define cell contents is less than n x m, then the grid will contain empty
cells.

If ORDER=ROWMAJOR and ROWS= is not defined, then the

number of rows is dynamically determined by the number of defined
cells.

ROWWEIGHTS=UNIFORM | PREFERRED | (weight-list)

specifies the proportional height of each row relative to the overall grid height, not
including the headings and sidebars.
UNIFORM
equally divides the total available height among all of the rows.
PREFERRED
specifies that each row gets its preferred height based on the following:
• Rows that contain one or more horizontally one-dimensional plots get the
maximum preferred height from the horizontally one-dimensional plots.
• The remaining rows that do not contain horizontally one-dimensional plots
get an equal amount of height from the remaining space.

Notes The PREFERRED option is used for lattice layouts that mix one-
dimensional and two-dimensional plots in the grid. It enables the layout
to compute the weights automatically for rows that contain one-
dimensional plots.

Examples of one-dimensional plots include axis tables, block plots,

fringe plots, and box plots with Y= values only.

(weight-list)
a space-separated list of row weights. The list should contain a weight for each
row, which can be expressed as one of the following:
PREFERRED
specifies that the corresponding row gets its preferred height as described
previously.

Note The PREFERRED option should be used on rows that contain only
one-dimensional plots. Using the PREFERRED keyword on rows that
contain a mix of one-dimensional and two-dimensional plots might
cause unexpected results. In that case, use a numeric weight instead.
LAYOUT LATTICE Statement 123

number
specifies that the corresponding row gets a height that is based on the
proportion of the specified number to the total of the numbers in the weight
list. For example, the following weight specifications are equivalent:
rowweights=(0.2 0.3 0.5)

rowweights=(2 3 5)

In these examples, the first row gets 20% of the available height, the second
row gets 30%, and the third row gets 50%.
If the list contains the PREFERRED keyword, then the number specifies the
proportion of the height that remains after the preferred height or heights are
calculated and subtracted from the available height.

Requirements The values in the weight list must be enclosed in parentheses.

If there are n rows in the grid, then the weight list should contain
n weights, one for each row.

All of the numbers in the list must be positive.

At least one number in the list must be nonzero.

Note The weights for all of the rows are specified as a proportion and,
as such, are not required to total 1.0.

Example Here is an example that specifies 25% of the available height for
the first row, 25% for the second row, and 50% for the third row.
rowweights=(1 1 2)

SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled, depending on the nesting levels of
the layouts that are used.

Default FALSE

Note Fonts maintain their size regardless of the specifications in the nested
layouts.

See “boolean ” on page 1339 for other Boolean values that you can use.

SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the unused cells in a partially filled lattice.
FALSE
displays external axes at their normal locations.
TRUE
skips empty cells and “snaps” the external axes to the nearest populated cell, both
vertically and horizontally.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

VALIGN=CENTER | TOP | BOTTOM | number

specifies this layout’s vertical alignment within its parent when nested within an
overlay-type or region layout.
124 Chapter 4 • Layout Statements

number
specifies the vertical alignment as a fraction of the parent container’s height.

Range 0 to 1, where 0 is on the bottom and 1 is on the top.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER

Restriction This option has effect only when this layout is nested within a region
layout, or when this layout is nested in an overlay-type layout and
AUTOALIGN=NONE.

See “LAYOUT OVERLAY Statement” on page 142 for more information

about how child positions are determined in an overlay-type or region
layout.

Example In the following example, the LATTICE layout is the child of the
OVERLAY layout. The child layout will appear in either the top right
or top left position, based on which position has more "unoccupied"
area.
dynamic VAR STAT1 STAT2 STAT3;
layout overlay / height=500px width=600px;
histogram VAR;
layout lattice / VALIGN=TOP HALIGN=RIGHT
height=80px width=70px columns=1;
entry STAT1;
entry STAT2;
entry STAT3;
endlayout;
endlayout;

SIDEBAR Optional Arguments

ALIGN=TOP | BOTTOM | LEFT | RIGHT
specifies the alignment of the sidebar.

Default BOTTOM

Tip You can use ENTRY statements to place rotated text in the right or left side
bar.

SPACEFILL=TRUE | FALSE
specifies whether to fill all the area of the sidebar with its contents.

Default TRUE

Tip To prevent a layout block within the sidebar from expanding to the sidebar
boundaries, set this option to FALSE.

See “boolean ” on page 1339 for other Boolean values that you can use.
LAYOUT LATTICE Statement 125

Details

Statement Description
The LAYOUT LATTICE statement creates a grid of graphs that are aligned across
columns and rows. For plot statements that are specified in the layout block or nested in
a LAYOUT OVERLAY statement, the LATTICE layout automatically aligns the plot
areas and tick display areas in the plots.
Note: To achieve the alignment, the LATTICE layout automatically aligns plot areas and
tick display areas across columns and rows. Also, it overrides axis-offset settings in
the OVERLAY layouts that you specify in those columns and rows. (For details
about offsets and the tick display area, see “Adjusting Axis Offsets” on page 886.) If
you do not want this alignment, then you might use LAYOUT GRIDDED instead.
For example, if you have a heterogeneous panel of graphs, such as a mix of scatter
plots, box plots, bar charts, or other types of graphs, then you might consider using
LAYOUT GRIDDED rather than LAYOUT LATTICE.
The layout can unify the scale of the data ranges that are displayed in the plots, based on
the values set for the COLUMNDATARANGE= and ROWDATARANGE= options. If
one or more plots within the template use the XAXIS= option to produce independent
X2 (top) axes, then the X2 data scales can be unified, based on the values set for the
COLUMN2DATARANGE= option. If one or more plots within the template use the
YAXIS= option to produce independent Y2 (right) axes, then the Y2 data scales can be
unified, based on the values set for the ROW2DATARANGE= options. The data ranges
can be scaled separately for each column, for each row, or for both. Or they can be
scaled across all columns, all rows, or all of both.
When the data-range scales are unified, you can simplify the layout by displaying only
the external axes that apply to all of the graphs across the corresponding columns or
rows. See “Axis Statements” on page 129 for more details.
The following figure shows the parts of the Lattice layout with the default axis display
(internal axes are displayed).
126 Chapter 4 • Layout Statements

This next figure shows the parts of the Lattice layout when the graph display is
simplified so that only external axes are displayed.
Note: The figure shows secondary X (top) and secondary Y (right) axes. The layout also
enables you to generate independent X2 (top) and independent Y2 (right) axes. For
details, see “Axis Statements” on page 129
LAYOUT LATTICE Statement 127

The columns and rows can be separated by areas called “gutters,” which are controlled
by the COLUMNGUTTER= and ROWGUTTER= options. In addition, the
COLUMNWEIGHTS= and ROWWEIGHTS= options can be used to allocate a
proportion of available space to each row and column.
The LATTICE layout automatically decides how much area to allocate to cell contents:
• Text items have a fixed size based on the amount of text and the font properties.
• Graphs take up the remaining space.
The layout’s grid size is determined by the COLUMNS= and ROWS= options.
By default, the results of the GTL-statements are placed into the grid sequentially from
left to right, wrapping to a new row each time the current row is filled. You can use the
ORDER= option to fill cells from top to bottom down a column. In that case, the layout
cells wrap to a new column each time the current column is filled.

Cell Contents
The following general syntax is used to define the contents of each cell in a LAYOUT
LATTICE:
GTL-statement(s) | cell-statement-block(s)
A cell-statement-block, when used, has the following syntax:
128 Chapter 4 • Layout Statements

CELL;
<CELLHEADER; GTL-statement(s); ENDCELLHEADER;>
GTL-statement(s);
ENDCELL;
The following guidelines apply to defining cell content:
• The contents of each cell is generated by GTL statements, which can be specified
independently or enclosed in a CELL block.
• Independent GTL statements include text statements, plot statements, or layout
blocks. Each independently specified GTL statement or layout block provides
content for one cell.
• A CELL block can include text statements, plot statements, or layout blocks. Each
CELL block provides content for one cell.
• Within a CELL block, you can use a CELLHEADER block to generate one or more
header lines within the cell. Specify each header line on a separate GTL statement
within the CELLHEADER block. The header block is typically used to specify one
or more text statements, but other statements are allowed within the block. For
example, you can specify a LAYOUT GRIDDED statement to produce a grid of text
for the header
• You can use only one CELLHEADER block per CELL block. If you specify more
than one, then only the last one is used.
• If you do not specify a CELLHEADER block in a CELL block, then the enclosed
GTL statements produce the same results that they would produce if they were
specified independently.
• If more than one plot statement is needed to generate contents for a cell, you should
place the plot statements in a layout block such as LAYOUT OVERLAY. Otherwise,
unexpected results might occur. This applies to independent GTL statements and to
GTL statements in a CELL block. See Figure 4.1 on page 129.
If a CELL block, or an independent statement or layout block for a lattice cell does not
produce any output, then the space for that cell might not be retained as an empty cell in
the lattice. In that case, the empty cell is removed, and the remaining cells (if any) fill
the gap in the lattice. A statement produces no output when required data for that
statement does not resolve. A layout block produces no output when it contains no
statements or when none of the statements contained in the block produce any output.
The example code shows a LAYOUT LATTICE block that uses one GTL statement and
one CELL block to generate the two-column layout shown in the following figure:
LAYOUT LATTICE Statement 129

Figure 4.1 Lattice Layout with Independent Plot Statements and a CELL Block

proc template;
define statgraph cellcontents;
begingraph;
layout lattice /
columngutter=5 columns=2;

/* independent plot statement - defines first cell */

scatterplot x=age y=height;

/* cell block - defines second cell */

cell;
cellheader;
entry "Cell Header" / border=true;
endcellheader;
/* two plot statements are needed - enclose
the statements in a LAYOUT OVERLAY block */
layout overlay;
scatterplot x=weight y=height;
referenceline y=53 / lineattrs=(pattern=shortdash)
curvelabellocation=inside curvelabel="Reference Line";
endlayout;
endcell;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.class template=cellcontents;

run;

Axis Statements

Overview
The axis statements can be used to simplify and clarify the layout by displaying only the
external axes in the resulting graph.
The following figure shows the default layout with internal axes displayed:
130 Chapter 4 • Layout Statements

This next figure shows a simplified layout with only the external axes displayed:

Axis statements are useful only if the data ranges across the affected columns or rows
are comparable and can be unified to a common scale. For example, external axes are
not supported if an affected lattice cell contains a LAYOUT OVERLAYEQUATED
statement. If the axis ranges are not unified for the affected columns or rows, then the
axis statements in the layout are ignored.
To unify data ranges in the layout grid, the following options are available:

Axis Option Axis Option

X COLUMNDATARANGE= X2 COLUMN2DATARANGE=

Y ROWDATARANGE= Y2 ROW2DATARANGE=
LAYOUT LATTICE Statement 131

Specifying Axis Features

For columns, axis features for the external X axes (bottom) are specified within a
COLUMNAXES block, nesting one COLUMNAXIS statement for each column that
contains an X axis that you need to manage. The COLUMNAXIS statement provides a
DISPLAYSECONDARY= option, which enables you to display a secondary X (top) axis
that mirrors the primary X axis but can have different display features. In that case, the
axis features that you specify in the COLUMNAXIS statement apply to both the primary
and secondary X axes.
If one or more plots within the template use the XAXIS= option to produce independent
X2 (top) axes, then axis features for the external X2 axes (top) are specified within a
COLUMN2AXES block, nesting one COLUMNAXIS statement for each column that
contains an X2 axis that you need to manage. Within the COLUMN2AXES block, the
COLUMNAXIS statement’s DISPLAYSECONDARY= option enables you to display a
secondary X2 (bottom) axis that mirrors the primary X2 axis but can have different
display features. Here again, the axis features that you specify in the COLUMNAXIS
statement apply to both the primary and secondary X2 axes.
Note: If you specify independent X and X2 scales at the same time, then the
DISPLAYSECONDARY= option is ignored in the COLUMNAXIS statement. This
is true whether the COLUMNAXIS statement is specified in a COLUMNAXES or
COLUMN2AXES block.
For both the COLUMNAXES and COLUMN2AXES blocks, if a lattice cell within the
column contains a LAYOUT OVERLAY with the XAXISOPTS= or X2AXISOPTS=
option specified, these OVERLAY options are ignored. In such cases, the desired axis
features should be specified in the COLUMNAXIS statement.
For rows, axis features for the external Y axes (left) are specified within a ROWAXES
block, nesting one ROWAXIS statement for each row that contains a Y axis that you
need to manage. The ROWAXIS statement provides a DISPLAYSECONDARY= option,
which enables you to display a secondary Y (right) axis that mirrors the primary Y axis
but can have different display features. In that case, the axis features that you specify in
the ROWAXIS statement apply to both the primary and secondary Y axes.
If one or more plots within the template use the YAXIS= option to produce independent
Y2 (right) axes, then axis features for the external Y2 axes (right) are specified within a
ROW2AXES block, nesting one ROWAXIS statement for each row that contains a Y2
axis that you need to manage. Within the ROW2AXES block, the ROWAXIS
statement’s DISPLAYSECONDARY= option enables you to display a secondary Y2
(left) axis that mirrors the primary Y2 axis but can have different display features. Here
again, the axis features that you specify in the ROWAXIS statement apply to both the
primary and secondary Y2 axes.
Note: If you specify independent Y and Y2 scales at the same time, then the
DISPLAYSECONDARY= option is ignored in the ROWAXIS statement. This is true
whether the ROWAXIS statement is specified in a ROWAXES or ROW2AXES
block.
For both the ROWAXES and ROW2AXES blocks, if a lattice cell within the row
contains a LAYOUT OVERLAY with the YAXISOPTS= or Y2AXISOPTS= option
specified, these OVERLAY options are ignored. In such cases, the desired axis features
should be specified in the ROWAXIS statement.

Syntax and Restrictions for the Axis Statements

The axis-statement blocks have the following general syntax:
132 Chapter 4 • Layout Statements

COLUMNAXES; COLUMN2AXES;
COLUMNAXIS / axis-option(s); COLUMNAXIS / axis-option(s);
<…COLUMNAXIS-n;> <…COLUMNAXIS-n;>
ENDCOLUMNAXES; ENDCOLUMN2AXES;

ROWAXES; ROW2AXES;
ROWAXIS / axis-option(s); ROWAXIS / axis-option(s);
<…ROWAXIS-n;> <…ROWAXIS-n;>
ENDROWAXES; ENDROW2AXES;

In the LATTICE layout block, the following restrictions apply:

• If the LAYOUT LATTICE statement sets the row or column data range to DATA,
then the corresponding axes block is ignored. The data range must be set to UNION
or UNIONALL to externalize the axes.
• Only one COLUMNAXES block can be used to manage X axes, and only one
COLUMN2AXES block can be used to manage X2 axes. If more than one of either
block is specified, then only the last one of that block type is used.
• Within a COLUMNAXES or COLUMN2AXES block, one COLUMNAXIS
statement should be specified for each column that contains axes that you need to
manage. Both axes blocks can contain a COLUMNAXIS statement for the same
column. For example, to manage the axes in the first column of the layout, the
COLUMNAXES block can contain a COLUMNAXIS statement that manages the
column’s X axes. The COLUMN2AXES block can contain a COLUMNAXIS
statement that manages the column’s X2 axes.
• Only one ROWAXES block can be used to manage Y axes, and only one
ROWAXES block can be used to manage Y2 axes. If more than one of either block
is specified, then only the last one of that block type is used.
• Within a ROWAXES or ROW2AXES block, one ROWAXIS statement should be
specified for each row that contains axes that you need to manage. Both axes blocks
can contain a ROWAXIS statement for the same row. For example, to manage the
axes in the first row of the layout, the ROWAXES block can contain a ROWAXIS
statement that manages the row’s Y axes. The ROW2AXES block can contain a
ROWAXIS statement that manages the column’s Y2 axes.
• If the number COLUMNLAXIS or ROWAXIS statements is greater than the number
needed, then the extra statements are ignored. If the number of statements is fewer
than the number needed, then the additional COLUMNAXIS or ROWAXIS
statements are automatically generated with DISPLAY=NONE options in effect.
For the list of axis-options, see “Axis Options for LAYOUT LATTICE” on page 963 .
The following example shows a LAYOUT LATTICE block that uses a ROWAXES
block to set external axes and display grid lines for the row display.
begingraph;
layout lattice /
rowdatarange=union
columns=2;

/* axis definitions */
rowaxes;
rowaxis /griddisplay=on;
endrowaxes;
LAYOUT LATTICE Statement 133

/* cell contents */
scatterplot x=x y=t;
scatterplot x=x y=y;

endlayout;
endgraph;

Here, the LAYOUT LATTICE statement specifies the ROWDATARANGE option to

unify the data ranges across rows in the layout. Because LAYOUT LATTICE specifies
COLUMNS=2 and there are two plot statements in the template, the resulting graph has
two columns and only one row. Thus, only one ROWAXIS statement is needed in the
ROWAXES block to specify axis attributes for that row of graphs. A ROW2AXES block
is not needed because neither SCATTERPLOT statement in the template maps data to
the Y2 axis.
For more information and examples that demonstrate how data are mapped to the axes,
see “Plot Data Are Mapped to a Designated Axis” on page 876 .

Header Statements
Header statements are used to display one or more headers for the columns and rows in a
Lattice layout. Each statement is specified as a block in the form statement -
ENDstatement. The header block is typically used to specify one or more text
statements, but other statements are allowed within the block. For example, you could
specify a LAYOUT GRIDDED statement to produce a grid of text for the header.
The general syntax for a COLUMNHEADERS statement is
COLUMNHEADERS;
GTL-statement(s);
ENDCOLUMNHEADERS;
The following header statements are available:

COLUMNHEADERS specifies a header for the primary (bottom) column-header

position.

COLUMN2HEADERS specifies a header for the secondary (top) column-header

position.

ROWHEADERS specifies a header for the primary (left) row-header position.

ENTRY statements can be used to specify rotated text.

ROW2HEADERS specifies a header for the secondary (right) row-header

position. ENTRY statements can be used to specify rotated
text.

• The LAYOUT LATTICE aligns headers with the columns, or the rows, or both.
• Each of the header blocks COLUMNHEADERS, COLUMN2HEADERS,
ROWHEADERS, and ROW2HEADERS can be used once in a LAYOUT LATTICE
block. If more than one block is specified for one of the statements, then only the last
specified block for that statement is used.
The following example shows a LAYOUT LATTICE block that uses a
COLUMNHEADERS block to display column headers above the left and right columns
in the layout.
134 Chapter 4 • Layout Statements

begingraph;
layout lattice / columns=2;

/* Lattice header definitions */

columnheaders;
entry "Left Column";
entry "Right Column";
endcolumnheaders;

/* cell contents */
scatter x=x y=t;
scatter x=x y=y;

endlayout;
endgraph;

Sidebar Statements
A LAYOUT LATTICE supports the display of a sidebar between a row or column
header and an external axis. (See the figures in “Statement Description” on page 125 .) A
sidebar spans across columns or rows and is useful for displaying information that
applies to all of the columns or all of the rows. For example, sidebars are useful for
displaying legends.
A SIDEBAR statement has the following syntax:
SIDEBAR </option(s)>;
GTL-statement(s);
ENDSIDEBAR;
• You can specify up to four SIDEBAR blocks in a LAYOUT LATTICE, one for each
of the top, bottom, left, and right sidebar positions.
• If two or more SIDEBAR blocks have the same alignment, then the sidebar
information forms two or more columns (ALIGN=LEFT or ALIGN=RIGHT) within
the sidebar area. Or it forms two or more rows (ALIGN=TOP or
ALIGN=BOTTOM) within the sidebar area.
• Only one statement (such as ENTRY or DISCRETELEGEND) or one layout block
(such as LAYOUT GRIDDED) is allowed in a SIDEBAR block. To create multi-line
text in a sidebar, nest ENTRY statements within a LAYOUT GRIDDED block.
• The LAYOUT LATTICE automatically aligns a sidebar with the layout columns or
rows.
The following example shows a LAYOUT LATTICE block that uses a SIDEBAR block
to display a top sidebar in the layout.
begingraph;
layout lattice / columns=2;

sidebar / align=top;
layout gridded / border=true ;
entry "Top Sidebar" ;
entry "(spans both columns)";
endlayout;
endsidebar;

scatterplot x=x y=t;

scatterolot x=x y=y;
Example: LAYOUT LATTICE Statement 135

endlayout;
begingraph;

Example: LAYOUT LATTICE Statement

This example shows a two-cell lattice layout (two columns, one row). A ROWAXES
block is used in the example to make the Y axes external to both cells.
• The ROWDATARANGE=UNION option assures that the data ranges of all Y=
columns in the row cells are considered to construct a common axis range. This
facilitates the visual comparison of the cells.
• A SIDEBAR block is used to place the legend at the top of the lattice.
• Because the ROWAXIS statement within the ROWAXES block uses the
DISPLAYSECONDARY= option, a secondary Y axis is displayed on the right. The
secondary Y axis is not an independent axis. Rather, it mirrors the primary Y axis,
making it easier to read Y-axis values when viewing the bar chart that is in the right
cell.
The following graph was generated by the “Example Program” on page 135 :

Example Program
proc template;
define statgraph layoutlattice;
begingraph;
entrytitle "Vehicle Gas Mileage";
entryfootnote "Averages of 428 models from 38 manufactures";
layout lattice / columns=2 rowdatarange=union;
136 Chapter 4 • Layout Statements

layout overlay / cycleattrs=true;

barchart category=origin response=mpg_highway /
stat=mean barwidth=0.8 name="H" ;
barchart category=origin response=mpg_city /
stat=mean barwidth=0.5 name="C" ;
endlayout;
layout overlay / cycleattrs=true;
barchart category=type response=mpg_highway /
stat=mean barwidth=0.8;
barchart category=type response=mpg_city /
stat=mean barwidth=0.5;
endlayout;
sidebar / align=top;
discretelegend "H" "C" / border=false;
endsidebar;
rowaxes;
rowaxis / display=(tickvalues)
displaysecondary=(tickvalues) griddisplay=on;
endrowaxes;
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.cars template=layoutlattice;
run;

LAYOUT OVERLAY Statement

Builds a composite from one or more GTL-statements. The composite could be an entire graph. Or, if this
layout is nested in a GRIDDED or LATTICE layout, then the composite typically provides contents for one
cell in the parent layout.
Restrictions: You can add one or more 2-D plots to the graph area that the LAYOUT OVERLAY
statement creates, but all of the graphs will share the same set of axes.
3-D plots are not allowed.
Interaction: When nested within another layout type, the OVERLAY layout defines the graph
display for one cell of the parent layout. A separate OVERLAY layout is specified for
each cell.

Syntax
LAYOUT OVERLAY </option(s)>;
GTL-statements;
<INNERMARGIN </options(s)>;
block-plot-statement(s); | axis-table statement(s);
ENDINNERMARGIN;>
<… more-innermargin-blocks …> >
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
LAYOUT OVERLAY Statement 137

ASPECTRATIO=AUTO | positive-number
specifies the aspect ratio of the plot’s wall area.
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in
nested plot statements automatically change from plot to plot.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
WALLDISPLAY=STANDARD | ALL | NONE | (display-options)
specifies whether the plot’s wall and wall outline are displayed.

Axes options
X2AXISOPTS=(axis-options)
specifies one or more X2 axis options.
XAXISOPTS=(axis-options)
specifies one or more X axis options.
Y2AXISOPTS=(axis-options)
specifies one or more Y2 axis options.
YAXISOPTS=(axis-options)
specifies one or more Y axis options.

Optional Arguments
ASPECTRATIO=AUTO | positive-number
specifies the aspect ratio of the plot’s wall area. The ratio is expressed as a positive
decimal fraction representing wall-height divided by wall-width. For example, 0.75
is a 3/4 aspect ratio and 1.0 is a square aspect ratio.

Default AUTO. The wall area is sized to the maximum area that can fill the
available space inside the OVERLAY layout.

Interaction When the LAYOUT OVERLAY statement is nested in a LAYOUT

LATTICE block, the ASPECTRATIO= option is ignored unless
ROWDATARANGE=DATA, COLUMNDATARANGE=DATA,
ROW2DATARANGE=DATA, and COLUMN2DATARANGE=DATA
are in effect in the LAYOUT LATTICE statement.

See “LAYOUT LATTICE Statement” on page 111

BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
138 Chapter 4 • Layout Statements

style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect for the color to be seen. By default,

OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in nested
plot statements automatically change from plot to plot.
FALSE
does not cycle the default visual attributes of multiple plots. For example, if you
overlay three series plots, then each series line has the same default visual
properties.
TRUE
attempts to use the GraphData1–GraphDataN style elements to assign different
visual properties to applicable plots (scatter plots and series plots and others).
Some plots in the layout do not participate in the cycling (for example, reference
lines and drop lines).

Default FALSE

See “Rotating Visual Attributes for Each Plot in an Overlay” on page 183

“boolean ” on page 1339 for other Boolean values that you can use.

Example In the following example, the first three series plots are assigned line
properties that are based on the GraphData1, GraphData2, and
GraphData3 style elements. The fourth series plot does not participate in
the attribute cycling because its LINEATTRS= option assigns a line style.
layout overlay / cycleattrs=true;
seriesplot x=date y=var1;
seriesplot x=date y=var2;
LAYOUT OVERLAY Statement 139

seriesplot x=date y=var3;

seriesplot x=date y=var4 / lineattrs=GraphReference;
endlayout;

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
140 Chapter 4 • Layout Statements

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style- attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.
STANDARD
displays a filled wall. The setting of the FRAMEBORDER= attribute of the
GraphWalls style element determines whether the wall outline is displayed.
ALL
displays a filled, outlined wall.
NONE
displays no wall and no wall outline.
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:

OUTLINE displays the wall outline.

FILL displays a filled wall area.
LAYOUT OVERLAY Statement 141

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall outline are set by the GraphAxisLine
style element.

XAXISOPTS=(axis-options)
specifies one or more X axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.

See “Axis Options for LAYOUT OVERLAY” on page 889 for a list of
axis options.

X2AXISOPTS=(axis-options)
specifies one or more X2 axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.

See “Axis Options for LAYOUT OVERLAY” on page 889 for a list of
axis options.

YAXISOPTS=(axis-options)
specifies one or more Y axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.

See “Axis Options for LAYOUT OVERLAY” on page 889 for a list of
options.

Y2AXISOPTS=(axis-options)
specifies one or more Y2 axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.

See “Axis Options for LAYOUT OVERLAY” on page 889 for a list of
options.
142 Chapter 4 • Layout Statements

Details
The LAYOUT OVERLAY statement builds a composite using one or more GTL-
statements. You can specify one or more two-dimensional plots within the layout,
provided all plots can share the same type of axes. You can also specify one or more
insets, such as nested layout statements (for example, LAYOUT GRIDDED), ENTRY
statements, and legend statements (for example, CONTINUOUSLEGEND or
DISCRETELEGEND).
The following general logic applies to rendering the composite:
Note: The details for positioning insets also apply to insets that are specified within a
LAYOUT REGION block.
• All plot statements are rendered first. Plot statement results are always rendered in
the plot area. The plots are stacked on top of one another in the order in which they
are specified, with the last one on top. It is possible for one plot’s graphical data to
obscure graphical data beneath it. You can control this by selectively ordering the
plot statements, or by using transparency on the individual plots, or by doing both.
• The insets are rendered next, in the order in which they are specified. As with the
plot statements, it is possible for the insets to obscure the results of other statements
in the layout.
• To control the horizontal and vertical positioning of some insets, you can use the
inset statement’s AUTOALIGN= option, or its HALIGN= and VALIGN= options.
Each nested inset determines its own relative position in the parent OVERLAY. This
positioning achieves the best results for text-based insets whose size can be easily fit
within an open area of the graph wall. A large text-based inset might not fit well, and
an inset that contains a plot might be dropped from the display without warning
when the template is executed.
• Some insets, like legends, can be positioned inside or outside of the plot area using
the inset statement’s LOCATION= option. The inset’s AUTOALIGN= or HALIGN=
and VALIGN= settings are then relative to that location.
Generally, the first specified plot determines the layout’s default axis characteristics. To
enable another plot to define the axis characteristics, set PRIMARY=TRUE for that plot.
For more information about the default axis characteristics, see “When Plots Share Data
and a Common Axis” on page 880. See also the LAYOUT OVERLAYEQUATED and
the LAYOUT OVERLAY3D statements.
An overlay layout can also contain an inner margin, which is a nested region at the top or
bottom of the OVERLAY container. One or more inner margin plots can be specified,
and each is specified within an INNERMARGIN block. Within the INNERMARGIN
block, only one-dimensional plot statements such as BLOCKPLOT and AXISTABLE
can be specified. See “INNERMARGIN Statement” on page 166.

Example: Simple Overlay

This example shows how to create a simple overlay of two series plots using the
OVERLAY layout. The following figure shows the output.
Example: Simple Overlay 143

Here is the SAS code.

data workers;
format Date monyy5.;
input Date monyy5. Electric Masonry;
datalines;
JAN80 320.3 293.8
FEB80 315.7 285.8
MAR80 312.6 292
APR80 306.5 299.3
MAY80 308.6 301.7
JUN80 316.3 307.9
JUL80 319.5 310.7
AUG80 326.4 314.9
SEP80 330.8 312.7
OCT80 329.3 318.5
NOV80 330.6 307.7
DEC80 327.2 296.2
JAN81 316.2 259.2
FEB81 310.1 258.8
MAR81 308.5 271.5
APR81 311.1 281
MAY81 313.6 283.7
JUN81 318.3 289.3
JUL81 321.3 291.1
AUG81 327.4 295.9
SEP81 326.7 292.7
OCT81 326.4 282.6
NOV81 322.5 275.5
DEC81 318.6 260.2
JAN82 301.9 214.3
FEB82 296.1 224.8
MAR82 298.3 228.7
144 Chapter 4 • Layout Statements

APR82 297.7 244.7

MAY82 303.5 260.4
JUN82 305 262.2
JUL82 307.6 270.4
;
proc template;
define statgraph layoutoverlay;
begingraph;
entrytitle "Trends in Employment Levels";
layout overlay / cycleattrs=true
xaxisopts=(display=(ticks tickvalues))
yaxisopts=(label="Number of Workers (thousands)");
seriesplot x=date y=electric /
curvelabel="Electrical"
curvelabellocation=outside;
seriesplot x=date y=masonry / curvelabel="Masonry"
curvelabellocation=outside;
endlayout;
endgraph;
end;
run;

proc sgrender data=workers template=layoutoverlay;

run;

LAYOUT OVERLAYEQUATED Statement

Builds a composite from one or more GTL-statements. The composite could be an entire graph. Or, if this
layout is nested in another layout, such as a GRIDDED layout, then the composite typically provides
contents for one cell in the parent layout. In an OVERLAYEQUATED layout, the display unit of the X axis
always equals the display unit of the Y axis.
Restrictions: All overlaid plots share common X and Y axes.
3-D plots are not allowed.
You can add one or more of the following X-Y plots to the graph area that the
LAYOUT OVERLAYEQUATED statement creates: BANDPLOT,
CONTOURPLOTPARM, ELLIPSE, ELLIPSEPARM, HEATMAP, HEATMAPPARM,
LOESSPLOT, NEEDLEPLOT, PBSPLINEPLOT, REGRESSIONPLOT,
SCATTERPLOT, SERIESPLOT, STEPPLOT, or VECTORPLOT. As long as one of
these plots is present, you can also add FRINGEPLOT, LINEPARM, MODELBAND,
REFERENCELINE, DROPLINE, DISCRETELEGEND, CONTINUOUSLEGEND, and
text-based statements such as ENTRY.
This layout has only two independent axes from a data standpoint, X and Y. If any
contained plot uses an X=X2 or Y=Y2 option, then the option is ignored and the data
is mapped to the X or Y axis. However, the X2 and Y2 axes can be displayed using
the DISPLAY2= suboption of the XAXISOPTS= and YAXISOPTS= options.
Interaction: When nested within another layout type, the OVERLAYEQUATED layout defines the
graph display for one cell of the parent layout. A separate OVERLAYEQUATED
layout is specified for each cell.
LAYOUT OVERLAYEQUATED Statement 145

Syntax
LAYOUT OVERLAYEQUATED </option(s)>;
GTL-statements;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in
nested plot statements automatically change from plot to plot.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
WALLDISPLAY=STANDARD | ALL | NONE | (display-options)
specifies whether the plot’s wall and wall outline are displayed.

Axes options
COMMONAXISOPTS=(common-axis-options)
specifies one or more options to apply to both the X and Y equated axes.
EQUATETYPE=FIT | SQUARE | SQUAREDATA | EQUATE
specifies how to draw the axis area.
XAXISOPTS=(axis-options)
specifies one or more X axis options.
YAXISOPTS=(axis-options)
specifies one or more Y axis options.

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

146 Chapter 4 • Layout Statements

Interaction OPAQUE=TRUE must be in effect for the color to be seen. By default,

OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

COMMONAXISOPTS=(common-axis-options)
specifies one or more options to apply to both the X and Y equated axes.

Requirements Axis options must be enclosed in parentheses.

Each option must be specified as a name = value pair and must be

separated by a space.

See “Options That Apply in Common to Both Equated Axes” on page

1013 for a list of options.

CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in nested
plot statements automatically change from plot to plot.
FALSE
does not cycle the default visual attributes of multiple plots. For example, if you
overlay three series plots, then each series line has the same default visual
properties.
TRUE
attempts to use the GraphData1–GraphDataN style elements to assign different
visual properties to applicable plots (scatter plots and series plots and others).
Some plots in the layout do not participate in the cycling (for example, reference
lines and drop lines).

Default FALSE

See “Rotating Visual Attributes for Each Plot in an Overlay” on page 183

“boolean ” on page 1339 for other Boolean values that you can use.

EQUATETYPE=FIT | SQUARE | SQUAREDATA | EQUATE

specifies how to draw the axis area.
LAYOUT OVERLAYEQUATED Statement 147

FIT
specifies that the X and Y axes have equal increments between tick values. The
data ranges of both axes are compared to establish a common increment size. The
axes might be of different lengths and have a different number of tick marks.
Each axis represents its own data range. One axis might be extended to use
available space in the plot area. If a TICKVALUELIST= or
TICKVALUESEQUENCE= axis option is used on COMMONAXISOPTS= ,
then it is ignored.
SQUARE
specifies that both the X and Y axes have the same length and the same major
tick values. The axis length and tick values are chosen so that the minimum and
maximum of both X and Y appear in the range of values appearing on both axes.
SQUAREDATA
specifies that both the X and Y axes have the same data range, but they can have
different tick values. A UNION of the data ranges does not occur in this case. For
example, if the X axis values are 20 to 40 (range of 20) and the Y axis values are
200 to 260 (range of 60), then both axes have a range of 60 units, but the X axis
can have tick values 0, 20, 40, and 60, and the Y axis can have tick values 200,
220, 240, and 260.
EQUATE
same as FIT except that neither axis is extended to use available space in the plot
area.

Default FIT

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
148 Chapter 4 • Layout Statements

AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.
LAYOUT OVERLAYEQUATED Statement 149

Default 0

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style- attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.
STANDARD
displays a filled wall. The setting of the FRAMEBORDER= attribute of the
GraphWalls style element determines whether the wall outline is displayed.
ALL
displays a filled, outlined wall.
NONE
displays no wall and no wall outline.
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:

OUTLINE displays the wall outline.

FILL displays a filled wall area.

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall outline are set by the GraphAxisLine
style element.

XAXISOPTS=(axis-options)
specifies one or more X axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.
150 Chapter 4 • Layout Statements

See “Options That Apply Separately to an X or Y Equated Axis” on

page 1019 for a list of options.

YAXISOPTS=(axis-options)
specifies one or more Y axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.

See “Options That Apply Separately to an X or Y Equated Axis” on

page 1019 for a list of options..

Details
The LAYOUT OVERLAYEQUATED statement is similar to the LAYOUT OVERLAY
statement: it builds a composite using one or more GTL-statements. Similar to a
LAYOUT OVERLAY, you can specify one or more 2-D plots within the layout,
provided all plots can share the same type of axes. (Additional restrictions are discussed
in a moment.) You can also specify one or more insets.
As an overlay-type layout, OVERLAYEQUATED has the same behavioral
characteristics as an OVERLAY layout. It uses the same general logic for rendering the
composite (see “LAYOUT OVERLAY Statement” on page 136 for details), and its
default axis characteristics are generally determined by the first specified plot, unless
you use PRIMARY=TRUE on an alternative plot statement (see “When Plots Share Data
and a Common Axis” on page 880).
OVERLAYEQUATED differs from OVERLAY in several ways. With
OVERLAYEQUATED,
• The axis type for both X and Y axes is always linear. Thus, plot types that have
discrete or binned axes cannot be used within this layout (for example, BOXPLOT,
BOXPLOTPARM, BARCHARTPARM, HISTOGRAM, and HISTOGRAMPARM).
• For equal data intervals on both axes, the display distance is the same. For example,
an interval of 2 on the X axis maps to the same display distance as an interval of 2 on
the Y axis.
• The aspect ratio of the plot display equals the aspect ratio of the plot data. In other
words, a 45 degree slope in data is represented by a 45 degree slope in the display.
The EQUATETYPE= option determines how the axes are drawn.
The following figure illustrates how a series plot might map differently when specified
in an OVERLAYEQUATED layout versus an OVERLAY layout:
Example: LAYOUT OVERLAYEQUATED Statement 151

A LAYOUT OVERLAYEQUATED statement enables you to specify one or more of the

following XY plots: SCATTERPLOT, SERIESPLOT, NEEDLEPLOT, STEPPLOT,
VECTORPLOT, BANDPLOT, LOESSPLOT, REGRESSIONPLOT, PBSPLINEPLOT,
and CONTOURPLOTPARM. As long as one of these plots is present, you can also add
FRINGEPLOT, LINEPARM, MODELBAND, REFERENCELINE, DROPLINE, and
insets as ENTRY, DISCRETELEGEND, and CONTINUOUSLEGEND.
From a data standpoint, this layout has only two independent axes, X and Y. If any plots
within the layout block use an XAXIS=X2 or YAXIS=Y2 option, then the option is
ignored and the data are mapped to the X or Y axis. To display X2 and Y2 axes, use the
DISPLAYSECONDARY= suboption of the XAXISOPTS= and YAXISOPTS= options.
If an OVERLAYEQUATED statement is nested in a LATTICE layout, then some of the
LATTICE’s alignment and external axis features are not supported on the
OVERLAYEQUATED layout.

Example: LAYOUT OVERLAYEQUATED Statement

The following graph was generated by the “Example Program” on page 152:
152 Chapter 4 • Layout Statements

Example Program
proc template;
define statgraph layoutoverlayequated;
begingraph;
entrytitle "Gas Mileage for GMC Models";
layout overlayequated / equatetype=fit;
referenceline y=16.2 /
curvelabel="City Average for Trucks/SUVs"
curvelabellocation=inside
curvelabelattrs=GraphReference;
referenceline x=20.6 /
curvelabel="Highway Average for Trucks/SUVs"
curvelabellocation=inside
curvelabelattrs=GraphReference;
scatterplot x=mpg_highway y=mpg_city /
datalabel=model;
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.cars
template=layoutoverlayequated;
where make="GMC";
run;

LAYOUT OVERLAY3D Statement

Builds a 3-D composite from one or more GTL-statements. The composite could be an entire graph. Or, if
this layout is nested in a GRIDDED or LATTICE layout, then the composite typically provides contents for
one cell in the parent layout.
LAYOUT OVERLAY3D Statement 153

Restriction: You can add one or more 3-D plots to the graph area that the LAYOUT OVERLAY3D
statement creates, but all of the graphs will share the same set of axes.

Syntax
LAYOUT OVERLAY3D </option(s)>;
GTL-statements;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CUBE=TRUE | FALSE
specifies whether the layout displays the lines that indicate the complete
bounding cube of the axis planes.
CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in
nested plot statements automatically change from plot to plot.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
ROTATE=number
specifies the angle of rotation.
TILT=number
specifies the angle of tilt in degrees.
WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
WALLDISPLAY=STANDARD | ALL | NONE | (display-options)
specifies whether the plot’s wall and wall outline are displayed.
ZOOM=positive-number
specifies a zoom factor.

Axes options
XAXISOPTS=(axis-options)
specifies one or more X axis options.
YAXISOPTS=(axis-options)
specifies one or more Y axis options.
ZAXISOPTS=(axis-options)
specifies one or more Z axis options.
154 Chapter 4 • Layout Statements

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect for the color to be seen. By default,

OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

CUBE=TRUE | FALSE
specifies whether the layout displays the lines that indicate the complete bounding
cube of the axis planes.

Default TRUE

Note The cube lines are displayed independently of the wall borders and axis
lines. Because some cube lines coincide with wall borders and axis lines, it
might appear that turning off wall borders or axis lines has no effect when
CUBE=TRUE.

Tip The color, thickness, and pattern of the cube lines are determined by the
GraphAxisLines style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in nested
plot statements automatically change from plot to plot.
LAYOUT OVERLAY3D Statement 155

FALSE
does not cycle the default visual attributes of multiple plots. For example, if you
overlay three series plots, then each series line has the same default visual
properties.
TRUE
attempts to use the GraphData1–GraphDataN style elements to assign different
visual properties to applicable plots (scatter plots and series plots and others).
Some plots in the layout do not participate in the cycling (for example, reference
lines and drop lines).

Default FALSE

See “Rotating Visual Attributes for Each Plot in an Overlay” on page 183

“boolean ” on page 1339 for other Boolean values that you can use.

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

156 Chapter 4 • Layout Statements

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

ROTATE=number
specifies the angle of rotation. Rotation is measured in a clockwise direction about a
virtual axis parallel to the Z axis (vertical) and passing through the center of the
bounding cube. A counterclockwise rotation can be specified with a negative value.

Default 54

TILT=number
specifies the angle of tilt in degrees. Tilt is measured in a clockwise direction about a
virtual axis parallel to the X axis (vertical) and passing through the center of the
bounding cube. A counterclockwise rotation can be specified with a negative value.

Default 20

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
LAYOUT OVERLAY3D Statement 157

style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style- attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.
STANDARD
displays a filled wall. The setting of the FRAMEBORDER= attribute of the
GraphWalls style element determines whether the wall outline is displayed.
ALL
displays a filled, outlined wall.
NONE
displays no wall and no wall outline.
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:

OUTLINE displays the wall outline.

FILL displays a filled wall area.

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall outline are set by the GraphAxisLine
style element.

See the CUBE= option.

XAXISOPTS=(axis-options)
specifies one or more X axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.

See “Axis Options for LAYOUT OVERLAY3D” on page 945 for a list
of axis options.

YAXISOPTS=(axis-options)
specifies one or more Y axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.
158 Chapter 4 • Layout Statements

See “Axis Options for LAYOUT OVERLAY3D” on page 945 for a list
of axis options.

ZAXISOPTS=(axis-options)
specifies one or more Z axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each pair
must be separated by a space.

See “Axis Options for LAYOUT OVERLAY3D” on page 945 for a list
of axis options..

ZOOM=positive-number
specifies a zoom factor. Factors greater than 1 move closer to the bounding cube, less
than 1 move farther away

Default 1

Details
The LAYOUT OVERLAY3D statement builds a 3-D composite using one or more GTL-
statements. You can specify one or more 3-D plots within the layout, provided all plots
can share the same type of axes. You can also specify “annotations” (for example, with
one or more ENTRY statements or LAYOUT GRIDDED statements). However,
annotations in the OVERLAY3D layout are more likely to collide with other graphics
features than are annotations in other overlay-type layouts.
As an overlay-type layout, OVERLAY3D has the same behavioral characteristics as an
OVERLAY layout. It uses the same general logic for rendering the composite (see
“LAYOUT OVERLAY Statement” on page 136 for details), and its default axis
characteristics are generally determined by the first specified plot, unless you use
PRIMARY=TRUE on another plot statement (see “When Plots Share Data and a
Common Axis” on page 880).
Within an OVERLAY3D layout, a graph’s bounding cube can be tilted, rotated, and
zoomed to provide a different viewpoint. By default, the outline of the bounding cube is
displayed and the viewing rotation angle is 57 degrees, the tilt angle is 20 degrees, and
the zoom factor is 1. See the CUBE= , ROTATE= , TILT= , and ZOOM= options for
information about how to change the viewpoint.

Example: LAYOUT OVERLAY3D Statement

The following graph was generated by the “Example Program” on page 159:
LAYOUT PROTOTYPE Statement 159

Example Program
proc template;
define statgraph layoutoverlay3d;
begingraph;
entrytitle "Density Plot of Height and Weight";
layout overlay3d / tilt=10 rotate=54
walldisplay=none cube=false;
surfaceplotparm x=height y=weight z=density /
surfacecolorgradient=density;
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.gridded template=layoutoverlay3d;
run;

LAYOUT PROTOTYPE Statement

Builds a composite from one or more plot-statements. The composite is used as a prototype or "rubber
stamp" that repeats in each cell of a parent DATALATTICE or DATAPANEL layout.
Restrictions: You can specify only one LAYOUT PROTOTYPE block in a LAYOUT DATAPANEL or
LAYOUT DATALATTICE block. If you specify more than one, then only the last
prototype block specified is honored. The remaining prototype blocks are ignored.
Only the following plots can be used in a LAYOUT PROTOTYPE block: BANDPLOT,
BARCHART, BARCHARTPARM, BLOCKPLOT, BOXPLOTPARM,
COUNTOURPLOTPARM, DROPLINE, ELLIPSEPARM, FRINGEPLOT,
HEATMAPPARM, HISTOGRAMPARM, LINECHART, LINEPARM, NEEDLEPLOT,
REFERENCELINE, SCATTERPLOT, SERIESPLOT, STEPPLOT, and
VECTORPLOT.
160 Chapter 4 • Layout Statements

SCATTERPLOTMATRIX plots, 3-D plots, and region plots such as PIECHART or

MOSAICPLOTPARM cannot be used in the LAYOUT PROTOTYPE block.
ENTRY, DISCRETELEGEND, and CONTINUOUSLEGEND statements cannot be
used in the LAYOUT PROTOTYPE block.
A plot statement cannot be used if it contains a column defined with an EVAL
expression.
You can add one or more two-dimensional plots and one-dimensional plots to the
graph area that the LAYOUT PROTOTYPE statement creates, provided all of the
graphs can share the same axis type.
If you include a plot statement with a CURVELABEL= option (such as
SERIESPLOT), then only CURVELABELLOCATION=INSIDE is supported.
If you include a plot statement that supports a CLIP= option (such as LINEPARM or
ELLIPSEPARM), then the CLIP value is always set to TRUE.
Requirement: The LAYOUT PROTOTYPE statement must be nested in a LAYOUT DATAPANEL or
LAYOUT DATALATTICE block.
Note: Nesting an INNERMARGIN block in the LAYOUT PROTOTYPE statement is valid in
the first maintenance release of SAS 9.4 and later releases.
See: “LAYOUT DATAPANEL Statement” on page 70
“LAYOUT DATALATTICE Statement” on page 45

Syntax
LAYOUT PROTOTYPE </option(s)>;
plot-statements;
<INNERMARGIN </options(s)>;
block-plot-statement(s); | axis-table statement(s);
ENDINNERMARGIN;>
<… more-innermargin-blocks …> >
ENDLAYOUT;

Optional Arguments
ASPECTRATIO=AUTO | positive-number
specifies the aspect ratio of the prototype cell. The ratio is expressed as a positive
decimal fraction representing wall-height divided by wall-width. For example, 0.75
is a 3/4 aspect ratio and 1.0 is a square aspect ratio.

Default AUTO. The prototype cell is sized to the maximum area that can fill the
available space inside the layout cell.

Note If AUTO is not used for the aspect ratio, then the entire DATALATTICE or
DATAPANEL grid is affected and changes shape.

CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in nested
plot statements automatically change from plot to plot.
FALSE
does not cycle the default visual attributes of multiple plots. For example, if you
overlay three series plots, then each series line has the same default visual
properties.
LAYOUT PROTOTYPE Statement 161

TRUE
attempts to use the GraphData1–GraphDataN style elements to assign different
visual properties to applicable plots (scatter plots and series plots and others).
Some plots in the layout do not participate in the cycling (for example, reference
lines and drop lines).

Default FALSE

See “Rotating Visual Attributes for Each Plot in an Overlay” on page 183

“boolean ” on page 1339 for other Boolean values that you can use.

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style- attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.
STANDARD
displays a filled wall. The setting of the FRAMEBORDER= attribute of the
GraphWalls style element determines whether the wall outline is displayed.
ALL
displays a filled, outlined wall.
NONE
displays no wall and no wall outline.
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:

OUTLINE displays the wall outline.

FILL displays a filled wall area.

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall outline are set by the GraphAxisLine
style element.

When the wall outline is suppressed, adjacent lines such as axis lines and
cell-header borders are still displayed. To suppress the axis lines, use the
appropriate display option for the axes. The cell-header borders cannot be
suppressed.
162 Chapter 4 • Layout Statements

Details
The LAYOUT PROTOTYPE statement defines a plot prototype or “rubber stamp” that
repeats automatically. It assembles the results of nested GTL statements into a single
axis area. The plots are drawn in the order in which they are specified. The results of the
last statement are placed on top.
The plot-statements determine the graphical content of the cells in the parent layout,
based on the subsetting of the specified classification variables. For an example, see
“LAYOUT DATALATTICE Statement” on page 45 or “LAYOUT DATAPANEL
Statement” on page 70.
A PROTOTYPE layout is essentially a restricted OVERLAY layout with the same
general rules for overlaying plots. The main difference is that there are no axis options
available on the LAYOUT PROTOTYPE statement. Axis properties are set with the
ROWAXISOPTS= and COLUMNAXISOPTS= options of the parent DATAPANEL or
DATALATTICE statement.
In SAS 9.4 and later releases, a PROTOTYPE layout can also contain an inner margin,
which is a nested region at the top or bottom of the PROTOTYPE container. One or
more inner margin plots can be specified, and each is specified within an
INNERMARGIN block. Within the INNERMARGIN block, only one-dimensional plot
statements such as BLOCKPLOT and AXISTABLE can be specified. See
“INNERMARGIN Statement” on page 166.

LAYOUT REGION Statement

Creates the drawing area for a plot that does not use axes.
Restrictions: A LAYOUT REGION block cannot contain more than one plot.
A LAYOUT REGION block can contain a PIECHART or MOSAICPLOTPARM plot
only.

Syntax
LAYOUT REGION </option(s)>;
GTL-statements;
ENDLAYOUT;

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect for the color to be seen. By default,

OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
LAYOUT REGION Statement 163

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the layout border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO
164 Chapter 4 • Layout Statements

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

Details
The REGION layout provides a container for plots that do not use axes. Within the
LAYOUT REGION block, you can specify a single plot statement of a type that never
uses axes, such as a PIECHART or MOSAICPLOTPARM. If multiple plot statements
are specified, then only the first one is honored. You can also specify one or more insets,
such as nested layout statements (for example, LAYOUT GRIDDED), ENTRY
statements, and legend statements (CONTINUOUSLEGEND or DISCRETELEGEND).
For example, you could specify a PIECHART statement with a DISCRETELEGEND
statement and an ENTRY statement. You can also nest one or more layout blocks within
the REGION layout. For example, you could nest a LAYOUT GRIDDED statement that
creates a small table of text.
Example: LAYOUT REGION Statement 165

When nested within another layout type, such as a GRIDDED or LATTICE layout, the
REGION layout defines the graphical display for one cell of the parent layout. A
separate REGION layout is specified for each cell.

Example: LAYOUT REGION Statement

The following graph was generated by the “Example Program” on page 165:

Example Program
proc template;
define statgraph layoutregion;
begingraph;
entrytitle "Average Weight by Age";
layout region;
piechart category=age response=weight /
stat=mean name="p"
datalabelcontent=(response) datalabellocation=outside;
discretelegend "p" / title="Age" across=2
border=true halign=right valign=top;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=layoutregion;

run;
166 Chapter 4 • Layout Statements

INNERMARGIN Statement
Provides a nested region in a LAYOUT OVERLAY or LAYOUT PROTOTYPE container in which a block
plot or axis table can be placed.
Restriction: This statement is valid in LAYOUT OVERLAY and LAYOUT PROTOTYPE blocks
only.
Notes: Two or more INNERMARGIN blocks that have the same alignment are stacked.
Multiple statements within an INNERMARGIN block are stacked.
For an X axis, the offsets on each end of the Y axis are increased to make room for
the inner margin plots. For a Y axis, the offsets on each end of the X axis are
increased to make room for the inner margin plots.

Syntax
INNERMARGIN < /option(s)>;
block-plot-statement(s); | axis-table statement(s);
ENDINNERMARGIN;

Optional Arguments
ALIGN=TOP | BOTTOM | LEFT | RIGHT
specifies the alignment of the inner margin.

Default BOTTOM

Restrictions For a block plot, only TOP and BOTTOM are valid. LEFT and
RIGHT are ignored.

For an axis table, LEFT and RIGHT can be used for a Y or Y2 axis.

Multiple statements within an INNERMARGIN block are stacked.

Note For an inner margin with ALIGN=TOP or ALIGN=BOTTOM, the

offsets on each end of the Y axis are increased to reserve space for the
inner margin plots. For an inner margin with ALIGN=LEFT or
ALIGN=RIGHT, the offsets on each end of the X axis are increased to
reserve space for the inner margin plots.
INNERMARGIN Statement 167

BACKGROUNDCOLOR=style-reference | color
specifies the color of the inner margin background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the style
attributes named COLOR or CONTRASTCOLOR are used.
color
specifies a color.

See color on page 1340

Default The graph wall color. (See WALLCOLOR=.)

Interaction For this option to have any effect, the OPAQUE= option must be set to
TRUE.

Note The inner margin background is set to the wall color even when
WALLDISPLAY= NONE.

GUTTER=dimension
specifies the gap between stacked items in the inner margin.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default 0

Requirement The inner margin must contain two or more stacked items for this
option to have any effect.

Note The default units for dimension are pixels.

See “dimension” on page 1340

OPAQUE=TRUE | FALSE
specifies whether the inner margin's background is opaque.

TRUE specifies that the background is opaque.

FALSE specifies that the background is transparent.

Default FALSE

Interaction When this option is FALSE, the BACKGROUNDCOLOR= option is

ignored.

Tip To prevent axis color bars and grid lines from passing through the axis
table, set OPAQUE=TRUE.

See “boolean ” on page 1339 for other Boolean values that you can use.

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the inner-margin border.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
168 Chapter 4 • Layout Statements

dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:
LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 5 px for the first inner margin adjacent to the bottom of the plot
area. Otherwise, 0.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 5 px for the first inner margin adjacent to the top of the plot area.
Otherwise, 0.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

SEPARATOR=TRUE | FALSE
specifies whether a separating line is drawn between the inner margin and the rest of
the layout content.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.

Default FALSE

Tip Use the SEPARATORATTRS= option to specify the attributes of the

separating line.

See “Example: Overlay with an Inner Margin Plot” on page 169

“boolean ” on page 1339 for other Boolean values that you can use.

SEPARATORATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the inner margin separating line.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
Example: Overlay with an Inner Margin Plot 169

Default The graphAxisLines style element

Interaction This option is ignored when SEPARATOR=FALSE.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

“Example: Overlay with an Inner Margin Plot” on page 169

Details
An inner margin is a nested region in an OVERLAY container. You can specify one or
more inner margin plots. Specify each plot within an INNERMARGIN block. Within an
INNERMARGIN block, you can specify only the BLOCKPLOT and AXISTABLE
statements. See “Example: Overlay with an Inner Margin Plot” on page 169.

Example: Overlay with an Inner Margin Plot

Example Overview
This example shows how to place a plot in an inner margin of an OVERLAY layout. It
creates a graph that shows monthly total sales for a specific year. A LINECHART
statement is used to draw the plot. The months are shown along the category axis, and
the total sales values are shown along the response axis. The Sashelp.Prdsale data set is
used as the data source. The tick marks on the category axis are positioned between the
midpoints to align with the beginning of each month.
A BLOCKPLOT statement in an INNERMARGIN block is used to display the quarters
above the category axis. The INNERMARGIN statement uses the default alignment, so
the inner margin is positioned at the bottom of the layout container, beneath the line
chart. Alternate shading is used in the block plot to show the block boundaries. Because
the tick marks are positioned between the midpoints, they align with the block
boundaries. The SEPARATOR= and SEPARATORATTRS= options are used in the
INNERMARGIN statement to specify a dark-red, two-pixel-wide separator line between
the inner margin and the rest of the graph.
Note: The INNERMARGIN statement SEPARATOR= and SEPARATORATTRS=
options are valid in the first maintenance release of SAS 9.4 and later releases.
170 Chapter 4 • Layout Statements

Example Output
Here is the output for this example.

Example Program
Here is the SAS code.
/* Create a format for the quarters */
proc format;
value quartername 1="Quarter 1" 2="Quarter 2"
3="Quarter 3" 4="Quarter 4";
run;

/* Define the graph template */

proc template;
define statgraph innermargin;
dynamic year;
begingraph / subpixel=on;
entrytitle "Total Sales in " year;
layout overlay /
xaxisopts=(type=discrete discreteopts=(ticktype=inbetween));
innermargin /
separator=true
separatorattrs=(color=darkred thickness=2px);
blockplot x=month block=quarter /
filltype=alternate
fillattrs=(color=cxd7d7d7)
altfillattrs=(color=cxf7f7f7)
display=(fill values) valuehalign=center;
endinnermargin;
linechart category=month response=actual /
smoothconnect=true;
endlayout;
endgraph;
end;
run;

/* Generate the graph */

proc sgrender data=sashelp.prdsale template=innermargin;
format quarter quartername.;
Example: Overlay with an Inner Margin Plot 171

where year=1994;
dynamic year=1994;
run;
172 Chapter 4 • Layout Statements
173

Part 4

Plot Statements

Chapter 5
Key Concepts for Using Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

Chapter 6
Plot Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
174
175

Chapter 5
Key Concepts for Using Plots

Minimum Requirements to Generate a Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

ODS Graphics Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Display Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Display Attributes for Non-Grouped Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Display Attributes for Grouped Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Rotating Visual Attributes for Each Plot in an Overlay . . . . . . . . . . . . . . . . . . . . . 183
Remapping Groups for Grouped Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Interactions between Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Location and Position of Curve Labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Curve Label Location Relative to the Plot Area . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Curve Label Position Relative to the Curve Line . . . . . . . . . . . . . . . . . . . . . . . . . . 186

Minimum Requirements to Generate a Plot

ODS graphics are generated by template definitions that determine a graph’s layout and
appearance and specify the variable roles to be represented in the graph display. A graph
can be rendered from a compiled template by associating the template with a data source
at run time.
The following SAS program shows the basic structure needed to meet the minimum
requirements for generating a plot using GTL:
proc template;
define statgraph minimumreq;
begingraph;
layout overlay;
scatterplot x=weight y=height;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.class template=minimumreq;

run;
176 Chapter 5 • Key Concepts for Using Plots

• The DEFINE STATGRAPH statement on PROC TEMPLATE is required to open a

definition block for defining and naming a graphics template. The END statement
closes the template definition.
• A BEGINGRAPH statement block is required to define the outermost container for
the graph. The ENDGRAPH statement closes the block.
• At least one layout statement block is required for specifying the elements that
compose the graph. To generate a plot, the layout block must contain at least one plot
statement. The ENDLAYOUT statement closes the layout block.
• The PROC TEMPLATE statement must be run to compile the template and save it in
the template store (Sasuser.Templat by default).
• The PROC SGRENDER statement is required to produce a graph from a compiled
template. The DATA= option specifies a run-time data source to use, and the
TEMPLATE= option specifies the template to use. The input data source must satisfy
any restrictions that are imposed by the template. For example, it must contain any
variables that have been specified on the template’s GTL statements.

ODS Graphics Environment

The ODS GRAPHICS statement manages the settings of the ODS Graphics environment
and is a statement that you will probably use frequently in your SAS sessions. For
example, the ODS GRAPHICS statement provides options that control the physical
aspects of your graphs, such as the image size and the name of the image file that is
created for the graph.
The default image size of 640 pixels by 480 pixels (4:3 aspect ratio) for ODS Graphics is
set in the SAS Registry. You can change the image size using the WIDTH= option, or
the HEIGHT= option, or both in the ODS GRAPHICS statement. To name the output
image file, use the IMAGENAME= option.
The following ODS GRAPHICS statement sets a 320 pixel width for the graph and
names the output image file modelfit:

ods graphics / width=320px

imagename="modelfit" reset;

proc sgrender data=sashelp.class template=modelfit;

run;

ods graphics off;

• The WIDTH= option sets the image width to 320 pixels. Because no HEIGHT=
option is used, SAS uses the design aspect ratio of the graph to compute the
appropriate height. (The width of 320px is half the default width, so SAS sets the
height to 240px, which is half the default height.)
• The IMAGENAME= option sets the name of the output image file to modelfit. The
RESET option ensures that each time the graph is produced, the previous version of
the image file is replaced. Otherwise, image names are incremented (modelfit1,
modelfit2, and so on) every time the graph is produced.
In general, it is good practice to specify only one sizing option without the other—just
the WIDTH= option or just the HEIGHT= option. That way SAS can maintain the
design aspect ratio of the graph, which might be important for many graphs. For
Display Attributes 177

example, a graph that has multiple columns or a statistics table on the side needs a wide
aspect ratio. Specifying both width and height in such cases might produce unpredictable
results.
Note: Size settings in the ODS GRAPHICS statement affect all of the graphs that are
rendered in the SAS session, unless they are changed by another ODS GRAPHICS
statement. The size for a graph produced by an individual template can be set with
the DESIGNWIDTH= and DESIGNHEIGHT= options in the BEGINGRAPH
statement. Size settings in the ODS GRAPHICS statement override size settings in
the BEGINGRAPH statement and remain in effect unless they are changed in
another ODS GRAPHICS statement or ODS GRAPHICS are turned off.
For more information about using the ODS GRAPHICS statement in GTL, see SAS
Graph Template Language: User's Guide. For a more complete discussion of the ODS
GRAPHICS statement, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.

Display Attributes

Overview
The display attributes for the lines, colors, marker symbols, and text used in a graph are
derived from the ODS style that is in effect when the graph is produced. These display
attributes might also be influenced by grouped data. To override default display
attributes, all GTL plot statements provide options that manage the graph’s visual
appearance. For example, a BOXPLOT statement provides an OUTLIERATTRS=
option that manages the visual appearance of outliers.
Two ways are generally available for modifying a graph’s display attributes:
• Change the ODS style that is in effect for the graph. “ODS Styles” on page 16
provides an overview of the use of styles in a graph. SAS Graph Template Language:
User's Guide discusses the use of styles in more detail.
• Override default style settings using GTL statement options. Some examples are
given in the sections that follow.

Display Attributes for Non-Grouped Data

Appendix 3, “Display Attributes,” on page 1347 documents the attribute settings that can
be specified for the lines, data markers, text, or area fills in a plot. The defaults for these
attributes are defined on style elements, but you can use attribute options on the plot
statement to change the defaults.
For example, the LINEPARM statement provides a LINEATTRS= option that specifies
the color, line pattern, or line thickness of the plot line. For non-grouped data, if you do
not set a line pattern in your template, then the default line pattern for the plot is
obtained from the GraphDataDefault:LineStyle style reference.
To change the default line pattern, a PATTERN= suboption on LINEATTRS= is
available. Figure 5.1 on page 178 shows the most common line patterns available for the
PATTERN= suboption.
178 Chapter 5 • Key Concepts for Using Plots

Figure 5.1 "Common Line Patterns"

• the left column shows the names for the line patterns
• the center column illustrates the type of line the name requests
• the right column shows the SAS line-style numbers for the line patterns
“Available Line Patterns” on page 1352 provides the complete list of line patterns that
can be used with the GTL.
In the following template definition, the LINEPARM statement’s LINEATTRS= option
overrides the default line pattern by specifying PATTERN=DASH:
proc template;
define statgraph patternchange;
begingraph;
layout overlay;
scatterplot y=height x=weight;
lineparm yintercept=intercept slope=slope /
lineattrs=(pattern=dash);
endlayout;
endgraph;
end;

Other display options can be managed the same way. For example, the SCATTERPLOT
statement provides a MARKERATTRS= option that specifies the color, size, symbol,
and weight of the plot data markers. For non-grouped data, if you do not set a marker
symbol in your template, then the default marker symbol is obtained from the
GraphDataDefault:MarkerSymbol style reference.
To change the default marker symbol, a SYMBOL= suboption on MARKERATTRS= is
available. Figure 5.2 on page 179 shows the marker symbols available for the
SYMBOL= suboption.
Display Attributes 179

Figure 5.2 "Marker Symbols"

In the following template definition, the SCATTERPLOT statement’s

MARKERATTRS= option overrides the default marker symbol by specifying
SYMBOL=CIRCLEFILLED, which uses a filled circle to represent the data points.
proc template;
define statgraph symbolchange;
begingraph;
layout overlay;
scatterplot y=height x=weight /
markerattrs=(symbol=circlefilled);
endlayout;
endgraph;
end;

Display Attributes for Grouped Data

Appendix 3, “Display Attributes,” on page 1347 documents the attribute settings that you
can specify for the lines, data markers, text, or area fills in a plot. For a grouped plot
(that is, when you use the GROUP= option in the plot statement), each distinct group
value can be represented in the graph by a different combination of line pattern, color,
and marker symbol (depending on the graph type). The defaults for these features are set
by the LineStyle, Color, ContrastColor, FillPattern, and MarkerSymbol attributes of the
GraphData1–GraphDataN style elements.
Note: The MarkerSize and LineThickness style attributes are not honored in the case of
grouped data.
When missing group values are displayed, the default attributes of the missing value are
set by the GraphMissing style element unless the MISSING= system option specifies a
character other than "." or " ". In that case, missing group value attributes are determined
by the GraphData1–GraphDataN style elements.
Figure 5.1 on page 178 shows the common line patterns available, and Figure 5.2 on
page 179 shows the marker symbols available.
For grouped plots, attributes such as colors, line patterns, and marker symbols are used
to distinguish the individual group values. The attributes are derived from the
GraphData1–GraphDataN style elements in the current style. The attributes are rotated
to provide distinct visual characteristics for each group value. For information about
attribute rotation, see “Attribute Rotation Patterns” in SAS Graph Template Language:
User's Guide. As discussed in “Rotating Visual Attributes for Each Plot in an Overlay”
on page 183, plot options might also influence the attribute rotation pattern.
180 Chapter 5 • Key Concepts for Using Plots

You can use attribute options on the plot statement to change the default display
attributes used for group data. For example, in the following template definition, the
LINEPARM statement’s LINEATTRS= option specifies PATTERN=DASH. This
explicit setting overrides the default line pattern for the plot lines and uses dashed lines
for all of the plots, leaving color to distinguish among group values.
proc template;
define statgraph dashedline;
begingraph;
layout overlay;
scatterplot y=height x=weight / group=gender;
lineparm yintercept=intercept slope=slope / group=gender
lineattrs=(pattern=dash);
endlayout;
endgraph;
end;

Rather than setting the same line pattern on all group values, you can change the default
sequence of line patterns that is used for grouped values. To do so, set the LineStyle
attribute in some of the style elements GraphData1–GraphDataN.
In the following example, a style is defined to change the line pattern for GraphData1
and GraphData2. In this example, the style is derived from the DEFAULT style. Values
are set for the LineStyle attributes in the GraphData1 and GraphData2 style elements.
The first default line in the sequence has long dashes (style value 6) and the second line
has short dashes (style value 4). The LineStyle settings for the remaining GraphData
elements are not set, so are derived from the parent style (DEFAULT). This new line
sequence is used as the default line sequence for any plot that uses the MyDefault style.
To apply the style to a graph, the STYLE= option is used in the ODS HTML statement
to specify the style name.
Here is the code for this example.
/* Specify a path for the ODS output */
filename odsout "output-path";

/* Sort the SASHELP.CLASS data by sex and age. */

proc sort data=sashelp.class(keep=height weight sex age)
out=class;
by sex age;
run;

/* Generate slope and intercept data for plot reference lines. */

proc robustreg data=class method=m
plots=none
outest=stats(rename=(weight=slope));
by sex;
model height=weight;
run;

data class;
merge class stats(keep=intercept slope sex);
run;

proc template;
/* Create custom style MYDEFAULT from the STYLES.DEFAULT style. */
define style MyDefault;
parent=Styles.Default;
Display Attributes 181

style GraphData1 from GraphData1 /

LineStyle=6;
style GraphData2 from GraphData2 /
LineStyle=4;
end;

/* Create the plot template. */

define statgraph testPattern;
begingraph;
layout overlay;
scatterplot y=height x=weight / group=sex;
lineparm x=0 y=intercept slope=slope / group=sex name="lines";
discretelegend "lines";
endlayout;
endgraph;
end;
run;

/* Generate the plot. */

ods _all_ close;
ods html path=odsout file="mydefaultstyle.html"
style=MyDefault; /* Apply style MyDefault to the graph. */

proc sgrender data=class template=testPattern;

run;

ods html close;

ods html; /* Not required in SAS Studio */

Here is the output.

Similarly, for grouped data, you can set the MarkerSymbol attribute in each of the style
elements GraphData1–GraphDataN. In the following example, a style is defined to
change the MarkerSymbol attribute for GraphData1–GraphData3. This new sequence is
used as the default marker symbol sequence for any grouped plot that uses the
MyDefault style.
Here is the code for this example.
182 Chapter 5 • Key Concepts for Using Plots

Note: The data that was generated in the previous example is used again in this
example.
/* Specify a path for the ODS output */
filename odsout "output-path";

proc template;
/* Create custom style MYDEFAULT from STYLES.DEFAULT. */
define style MyDefault;
parent=Styles.Default;
style GraphData1 from GraphData1 /
MarkerSymbol="DIAMOND";
style GraphData2 from GraphData2 /
MarkerSymbol="CROSS";
style GraphData3 from GraphData3 /
MarkerSymbol="CIRCLE";
end;

/* Create the plot template. */

define statgraph testSymbols;
begingraph;
layout Overlay;
scatterPlot y=height x=weight / group=age name="symbols";
discretelegend "symbols" / title="Age";
endlayout;
endgraph;
end;
run;

/* Generate the plot. */

ods _all_ close;
ods html path=odsout file="mydefaultstyle.html"
style=MyDefault; /* Apply style MyDefault to the graph. */

proc sgrender data=class template=testSymbols;

run;
Display Attributes 183

Rotating Visual Attributes for Each Plot in an Overlay

Overlay-type layouts provide the CYCLEATTRS= option, which specifies whether the
default visual attributes of lines, marker symbols, and area fills in nested plot statements
automatically change from plot to plot. When CYCLEATTRS=TRUE, all applicable plot
statements (SCATTERPLOT, SERIESPLOT, and others) are sequentially assigned the
next unused GraphDataN style element. (The sequence is overridden for plot statements
that have an explicit setting, either through a style element assignment or option
settings.) No plot retains its default (implicit) style element.
In the following example, assuming ungrouped data and the default attribute rotation
pattern, the series plots are assigned line properties based on the GraphData1,
GraphData2, and GraphData3 style elements. The reference line uses GraphReference,
not GraphData4.

layout overlay / cycleattrs=true;

seriesplot x=date y=var1;
seriesplot x=date y=var2;
seriesplot x=date y=var3;
referenceline x=cutoff / lineattrs=GraphReference;
endlayout;

If one of the plots in this example uses grouped data, then the grouped plots also
participate in the default cycles. For example, if the second plot has three groups, then it
generates three plots, which are assigned line properties based on the GraphData2,
GraphData3, and GraphData4 style elements.
If the plot statement that uses grouped data also uses the INDEX= option to manage the
group values (see “Remapping Groups for Grouped Data” on page 183), then the
INDEX= option overrides the default behavior. In that case, the grouped plots do not
participate in the default cycling.
When one or more of the plots within the layout override the default cycling behavior,
the arrangement of the plots within the layout might affect the default mapping of the
GraphDataN elements to those statements that participate in the default cycling.

Remapping Groups for Grouped Data

Indexing can be used to collapse the number of groups that are represented in a graph.
For example, if 10 groups are in the data, then indexes 1 and 2 can be assigned to the
first two groups, and index 3 can be assigned to all other groups. The third through tenth
data groups are treated as a single group in the graph.
Indexing can control the order in which colors, area fills, marker symbols, and line styles
are mapped to group values in a graph. This ordering method is needed only for
coordinating the data display of multiple graphs when the default mapping would cause
group values to be mismatched between graphs.
For example, consider two studies of three drugs, A, B, and C. If Study 1 uses all three
drugs, then the first combination of color and marker symbol is mapped to Drug A. The
second combination of color and marker symbol is mapped to Drug B, and the third is
mapped to Drug C. If Study 2 omits Drug A, then the first combination of color and
marker symbol is mapped to Drug B, and the second is mapped to Drug C. If the two
graphs are viewed together, then this default mapping causes the group values to be
mismatched. The visual attributes that represent Drug A in the first graph represent Drug
184 Chapter 5 • Key Concepts for Using Plots

B in the second graph. Those that represent Drug B in the first graph represent Drug C in
the second group.
The GROUP= option mappings can be made consistent between the two graphs by
creating an index column for each study. For these example studies, the GROUP and
INDEX columns are the following:

Table 5.1 Study 1

Drug1 Index1

A 1

B 2

B 2

C 3

Table 5.2 Study 2

Drug2 Index2

B 2

C 3

C 3

If the graph for Study 1 specifies INDEX=INDEX1 and the graph for Study 2 specifies
INDEX=INDEX2, then the second combination of color and marker symbol is mapped
to Drug B in both graphs. The third combination of color and marker symbol is mapped
to Drug C in both graphs.

Interactions between Options

When you use GTL statement options to manage the graph display, interactions between
options might cause some option settings to be ignored. For example, an ENTRYTITLE
statement provides BORDER= and BORDERATTRS= options for managing a border
line around the graph title. Border attributes that are set on the BORDERATTRS= option
have no effect on the graph title unless the title border line is displayed by setting
BORDER=TRUE.
Similarly, if a BOXPLOT statement’s DISPLAY= option suppresses the display of
outliers in a box plot, then using the OUTLIERATTRS= option to set outlier attributes
has no effect. The OUTLIERATTRS= settings only take effect if DISPLAY= enables the
display of outliers.
The option interactions are not limited to options that simply manage visual elements.
For example, on a BOXPLOT, if the EXTREME= option extends the box whiskers
beyond the fences, then outliers are suppressed in the plot and options that affect the
outliers are ignored, if set.
Location and Position of Curve Labels 185

The documentation for each GTL statement identifies the option interactions that might
occur on that statement.

Location and Position of Curve Labels

Overview
On plots that generate a curve line (a series plot or a density plot, for example), you can
specify a label for the curve line. You can also determine the label’s location in the
graph. For example, the SERIESPLOT statement provides the following options for
managing a curve label:
CURVELABEL
specifies a label for the curve line.
CURVELABELLOCATION
specifies the location of the curve line label relative to the plot area.
CURVELABELPOSITION
specifies the position of the label relative to the curve line.

Curve Label Location Relative to the Plot Area

By default, the label for a curve line is displayed inside the plot area. The following
figure shows the default location of the label for a series plot labeled “Curve Label”:

Depending on the shape of the curve line, its distribution of values, and the other plot
elements that must be displayed within the plot area, GTL might have to add an offset
(see “Adjusting Axis Offsets” on page 886) to one of the plot’s axis lines to provide
enough room for the curve label. To prevent the offset of the axis line, you can move the
curve label outside of the plot area by specifying
CURVELABELLOCATION=OUTSIDE on the plot statement:
186 Chapter 5 • Key Concepts for Using Plots

Regardless of whether the curve label is displayed inside or outside of the plot area, you
can use the CURVELABELPOSITION= option to adjust the label’s position relative to
the curve line.

Curve Label Position Relative to the Curve Line

Given a curve label’s location inside or outside of the plot area, a plot statement’s
CURVELABELPOSITION= option can adjust the label’s position relative to the curve
line. For example, the following positions are available for a series plot (for some plots,
START and END are not available):
AUTO
positions the curve label automatically near the end series line along unused axes
whenever possible (typically Y2 or X2) to avoid collision with tick values. This
position is used only when CURVELABELLOCATION=OUTSIDE.
MAX
forces the curve label to appear near maximum series values (typically, to the right).
MIN
forces the curve label to appear near minimum series values (typically, to the left).
START
forces the curve label to appear near the beginning of the curve. This position is
particularly useful when the curve line has a spiral shape. It is used only when
CURVELABELLOCATION=INSIDE.
END
forces the curve label to appear near the end of the curve. This position is
particularly useful when the curve line has a spiral shape. It is used only when
CURVELABELLOCATION=INSIDE.
When CURVELABELLOCATION=INSIDE, you can choose whether to position the
curve label near the START or END of the curve, or near the minimum data values
(MIN) or maximum data values (MAX). START and END use a different algorithm than
MIN and MAX. They are particularly useful for spiral-shaped curves whose end points
do not correlate with the minimum and maximum data values. In those cases, START or
END provide “better” label locations than MIN and MAX.
When CURVELABELLOCATION=OUTSIDE and CURVELABELPOSITION=AUTO,
a “good” position is automatically chosen to avoid collision with the axis information.
The following figure shows the different combinations of label locations and positions:
Location and Position of Curve Labels 187

• The minimum or maximum axis tick marks can be adjusted (see “Adjusting Axis
Offsets” on page 886) so that the label can be placed inside the plot area. Increasing
label length decreases the area available for displaying plots.
• When CURVLABELLOCATION=OUTSIDE, you can set the
CURVELABELPOSITION to MIN or MAX, but the label might collide with the
axis ticks and tick values, unless you are aware of where the axes are positioned.
188 Chapter 5 • Key Concepts for Using Plots
189

Chapter 6
Plot Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
AXISTABLE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
BANDPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
BARCHART Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
BARCHARTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
BIHISTOGRAM3DPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
BLOCKPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
BOXPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
BOXPLOTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
BUBBLEPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
CONTOURPLOTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
DENDROGRAM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
DENSITYPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
DROPLINE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
ELLIPSE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
ELLIPSEPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
FRINGEPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
HEATMAP Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
HEATMAPPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
HIGHLOWPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471
HISTOGRAM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
HISTOGRAMPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
LINECHART Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
LINEPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542
LOESSPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
MODELBAND Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565
MOSAICPLOTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573
NEEDLEPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584
PBSPLINEPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 600
PIECHART Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613
POLYGONPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 628
REFERENCELINE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 654
REGRESSIONPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 666
SCATTERPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 679
SCATTERPLOTMATRIX Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 715
SERIESPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 740
STEPPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 774
SURFACEPLOTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 803
TEXTPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 811
VECTORPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 837
WATERFALLCHART Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854
190 Chapter 6 • Plot Statements

Dictionary

AXISTABLE Statement
Creates an event plot of input data along an axis of an X-Y plot.

Syntax
AXISTABLE X=column | expression VALUE=column </option(s)>;
AXISTABLE Y=column | expression VALUE=column </option(s)>;

Summary of Optional Arguments

Appearance options
CLASS=column | expression
creates a separate row or column for each unique class value.
CLASSORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the order in which the class values are displayed.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing
or bin width.
COLORGROUP=column | expression | discrete-attr-var
specifies a column that is used to discretely map the color of the value text.
DATATRANSPARENCY=number
specifies the degree of the transparency of the header, label, and values.
DISPLAY=STANDARD | ALL | (display-options)
specifies which features to display.
DROPONMISSING=TRUE | FALSE
specifies whether the entire axis table is dropped when all of the VALUE=
column values are missing.
GUTTER=dimension
specifies the gap between rows when a class variable is used.
INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether missing class values are represented in the table.
INDENT=dimension
specifies a value to be used with the INDENTWEIGHT= option to determine
the indention for each text value.
INDENTWEIGHT=numeric-column | expression
specifies the indention weight (multiplier) for each observation.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the table border.
POSITION=number
positions the plot along the axis orthogonal to the axis used for the values.
SHOWMISSING=TRUE | FALSE
specifies whether missing values are represented in the table.
AXISTABLE Statement 191

TEXTGROUP=discrete-attr-var
specifies the discrete attribute variable for a discrete attribute map that maps
text attributes to values for each observation.
VALUEATTRS=style-element | style-element(text-options) | (text-options)
specifies the color and font attributes of the text values.
VALUEFORMAT=format
specifies a SAS format or a user-defined format for the table values.
VALUEHALIGN=AUTO | LEFT | CENTER | RIGHT
in a Y-axis table, specifies the horizontal alignment of the column values
relative to the column width.
VALUEJUSTIFY=AUTO | LEFT | CENTER | RIGHT
specifies the justification of the values in the axis table.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Header options
HEADERLABEL="string"
specifies the text for the table header.
HEADERLABELATTRS=style-element | style-element(text-options) | (text-options)
specifies the color and font attributes of the table header.
TITLE="string"
specifies the text for the table title.
TITLEATTRS=style-element | style-element(text-options) | (text-options)
specifies the color and font attributes of the table title.
TITLEHALIGN=AUTO | CENTER | LEFT | RIGHT
specifies the horizontal alignment of the axis table header label relative to the
axis table width.
TITLEJUSTIFY=LEFT | CENTER | RIGHT
specifies the justification of the title string. The justification is relative to the
axis table width.

Label options
LABEL="string"
specifies the text for the table label.
LABELATTRS=style-element | style-element(text-options) | (text-options)
specifies the color and font attributes of the column label.
LABELHALIGN=AUTO | LEFT | CENTER | RIGHT
specifies the horizontal alignment of the column label when it is displayed.
LABELJUSTIFY=LEFT | CENTER | RIGHT
specifies the justification of the column label when it is displayed.
LABELPOSITION=MIN | MAX
specifies the end of the axis on which the label is displayed.

Midpoint options
CLASSDISPLAY=STACK | CLUSTER
192 Chapter 6 • Plot Statements

specifies how the class values are displayed.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
STAT=AUTO | SUM | MEAN
specifies the statistic that is to be computed for the VALUE= column when
the column is numeric.

Required Arguments
Either the X= or the Y= argument must be specified in the AXISTABLE statement.
Specifying X= places an axis table along the X axis of a plot. Specifying Y= places an
axis table along the Y axis of a plot.
X=column | expression
specifies the column for the X axis.

Requirement If not specified, then Y= must be specified.

Y=column | expression
specifies the column for the Y axis.

Requirement If not specified, then X= must be specified.

VALUE=column
specifies the column that contains the axis table values.

Optional Arguments
CLASS=column | expression
creates a separate row or column for each unique class value. Each row or column is
labeled by the class value.

Interaction The DISPLAY= option that is in effect must include LABEL for any
labels to appear.

Tip Use the LABELATTRS= option to modify the label text attributes.

CLASSDISPLAY=STACK | CLUSTER
specifies how the class values are displayed.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
STACK
displays the class values vertically at each midpoint value on the X axis or
horizontally on the Y axis.
CLUSTER
displays the class values horizontally at each midpoint value on the X axis or
vertically on the Y axis.

Restriction CLUSTER applies only when the axis table is on a discrete axis.

Tip The CLUSTERWIDTH= option controls the cluster width.

AXISTABLE Statement 193

Default STACK

Interaction The CLASS= option must be specified for this option to have any
effect.

CLASSORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the order in which the class values are displayed.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
DATA
displays the class values in the order in which they occur in the data.
REVERSEDATA
displays the class values in the reverse order from which they occur in the data.

Tip This option is useful when the plot axis is reversed.

ASCENDING | DESCENDING
displays the class values in ascending or descending order.

Default DATA

Interaction The CLASS= option must be specified for this option to have any
effect.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or bin
width.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.

Requirement For this option to take effect, the CLASS= option must also be
specified, and the CLASSDISPLAY= option must be set to
CLUSTER.

COLORGROUP=column | expression | discrete-attr-var

specifies a column that is used to discretely map the color of the value text.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Each unique value of this column is mapped to the COLOR attribute of the
GraphData1–GraphDataN style elements that are in effect. If a discrete attribute
variable is specified, the color mapping from its associated DISCRETEATTRMAP
statement is used.

Interaction This option is ignored when the TEXTGROUP= option is specified.

194 Chapter 6 • Plot Statements

See “DISCRETEATTRVAR Statement” on page 1297

“DISCRETEATTRMAP Statement” on page 1287

DATATRANSPARENCY=number
specifies the degree of the transparency of the header, label, and values.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISPLAY=STANDARD | ALL | (display-options)

specifies which features to display.
STANDARD
displays the table values and, if provided, the table label.
ALL
displays the same features as STANDARD.
(display-options)
a space-separated list of display options, enclosed in parentheses. The following
options are supported:
LABEL
displays the table label. The label can be the VALUE= column label or name,
the LABEL= value, or the CLASS= value for the table, depending on the
options that you specify.
VALUES
displays the column values.
Note: This feature applies to the second maintenance release of SAS 9.4 and
to later releases.

Tip The column values are always displayed, even if DISPLAY=(LABEL) is

specified. To hide the table label, specify DISPLAY=(VALUES).

Default STANDARD

Note If a table title is specified, it is always displayed.

DROPONMISSING=TRUE | FALSE
specifies whether the entire axis table is dropped when all of the VALUE= column
values are missing.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default FALSE

Tip The SHOWMISSING= option controls whether missing values are shown
in the table.

See VALUE= on page 192

“boolean ” on page 1339 for other Boolean values that you can use.

GUTTER=dimension
specifies the gap between rows when a class variable is used.
AXISTABLE Statement 195

Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Defaults Y-axis table: 8 px

X-axis table: 0 px

Interaction The CLASS= option must be specified for this option to have any
effect.

See “dimension” on page 1340

HEADERLABEL="string"
specifies the text for the table header.
Note: Starting with the second maintenance release of SAS 9.4, the
HEADERLABEL= option is deprecated and is replaced with the TITLE= option.
The syntax and functionality are the same. The HEADERLABEL= option is still
honored, but the TITLE= option is preferred.

Default No table header is displayed

Tip Use the HEADERLABELATTRS= option to control the appearance of the

table header.

HEADERLABELATTRS=style-element | style-element(text-options) | (text-options)

specifies the color and font attributes of the table header.
Note: Starting with the second maintenance release of SAS 9.4, the
HEADERLABELATTRS= option is deprecated and is replaced with the
TITLEATTRS= option. The syntax and functionality are the same. The
HEADERLABELATTRS= option is still honored, but the TITLEATTRS= option
is preferred.

See “TITLEATTRS=style-element | style-element(text-options) | (text-options)” on

page 200

INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether missing class values are represented in the table.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
Missing class values are included by default. When the data contains missing class
values, the label for those values is either blank for missing character values or a dot
for missing numeric values.
The following figure shows an X-axis axis table that displays values for classes Class 1,
Class 2, and any missing class values.

Notice that the label for the missing class values is blank. You can use the
INCLUDEMISSINGCLASS=FALSE option to exclude the missing class values. If
you want to keep the missing class values, then you can create a format that specifies
196 Chapter 6 • Plot Statements

a more meaningful label for the missing class. For example, here is a format that
specifies a label for missing character and numeric class values.
proc format;
value $missingClass " " = "(Missing)";
value missingClass . = "(Missing)";
run;

A single space enclosed in quotation marks specifies a missing character value and a
dot specifies a missing numeric value. Although it might seem appropriate to use
empty quotation marks ('' or "") to specify a missing character value, doing so
produces unexpected results. To specify a missing character value, enclose a single
space in quotation marks (' ' or " "). You can use this format for the class columns in
the PROC SGRENDER statement. In that case, if the class columns contain missing
values, then the labels specified in the format statement are used for the missing
classes.
The following figure shows the previous example when format $missingClass is
applied to the class variable.

Note: In the second maintenance release of SAS 9.4 and in earlier releases, ODS
Graphics does not support Unicode values in user-defined formats. Starting with
the third maintenance release of SAS 9.4, ODS Graphics supports Unicode
values in user-defined formats only if they are preceded by the (*ESC*) escape
sequence. Example: "(*ESC*){unicode beta}". ODS Graphics does not
support an escape character that is defined in an ODS ESCAPECHAR statement
in user-defined formats.

Default TRUE

Interaction The CLASS= option must be specified for this option to have any
effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDENT=dimension
specifies a value to be used with the INDENTWEIGHT= option to determine the
indention for each text value.

Default 1/8 inch

Interaction The INDENTWEIGHT= option must be specified for this option to

have any effect.

See “dimension” on page 1340

INDENTWEIGHT=numeric-column | expression
specifies the indention weight (multiplier) for each observation.
AXISTABLE Statement 197

Interaction For each observation, the INDENT= option value is multiplied by the
value of the column specified by this option to determine the indention
for that observation’s value.

LABEL="string"
specifies the text for the table label.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default The class values are used when the CLASS= option is set and
CLASSDISPLAY=STACK is in effect. Otherwise, the VALUE=
column label or name is used.

Interaction This option is ignored when the CLASS= option is in effect.

Note If the length of the label exceeds the available space, the label is split
on blank space as needed to make it fit.

See CLASS= on page 192

LABELATTRS=style-element | style-element(text-options) | (text-options)

specifies the color and font attributes of the column label.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, the label color changes to match the group color
derived from the ContrastColor attribute of the GraphData1–
GraphDataN style elements.

Restriction Group behavior occurs only when the CLASS= and COLORGROUP=
option values are the same.

Interaction If one or more text options are specified and they do not include all the
font properties such as color, family, size, weight, and style, then the
properties that are not specified are derived from the GraphValueText
style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LABELHALIGN=AUTO | LEFT | CENTER | RIGHT

specifies the horizontal alignment of the column label when it is displayed.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
The alignment is relative to the column width.
AUTO
uses the effective value of the LABELJUSTIFY= option.
LEFT | CENTER | RIGHT
horizontally justifies the label left, center, or right,

Default AUTO
198 Chapter 6 • Plot Statements

Restriction This option applies only to Y-axis tables.

Interaction The DISPLAY= option must include LABEL for this option to have
any effect.

LABELJUSTIFY=LEFT | CENTER | RIGHT

specifies the justification of the column label when it is displayed.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
The justification is relative to the column width.

Default CENTER

Restriction This option applies only to Y-axis tables.

Interaction The DISPLAY= option must include LABEL for this option to have
any effect.

LABELPOSITION=MIN | MAX
specifies the end of the axis on which the label is displayed. The label is aligned with
the tick values on the axis.

Default MIN

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the table border.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the table border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options,
enclosed in parentheses:
LEFT=dimension
specifies the amount of extra space added to the left side.

Default 4 px

Restriction This option applies only to Y-axis tables.

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 4 px
AXISTABLE Statement 199

Restriction This option applies only to Y-axis tables.

TOP=dimension
specifies the amount of extra space added to the top.

Default 0 px

Restriction This option applies only to X axis tables.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0 px

Restriction This option applies only to X axis tables.

Note Sides that are not assigned padding are padded with the default amount of
space.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

POSITION=number
positions the plot along the axis orthogonal to the axis used for the values. This
option enables you to position the plot when the AXISTABLE statement is not
placed in an INNERMARGIN block.
number
specifies the position on the orthogonal axis as a fraction of the axis range.

Default Determined by the system.

Range 0 (bottom)–1 (top)

Interaction This option is ignored when the AXISTABLE statement is placed in an

INNERMARGIN block. It is also ignored when the AXISTABLE
statement is placed in a LAYOUT OVERLAY block by itself.

Tip To reserve space for the plot at either end of the orthogonal axis,
specify the OFFSETMIN= or OFFSETMAX= option for the orthogonal
axis.

SHOWMISSING=TRUE | FALSE
specifies whether missing values are represented in the table.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
The values are evaluated before the column format is applied. When this option is set
to FALSE, missing numeric and character values are hidden.

Default TRUE

See “boolean ” on page 1339 for other Boolean values that you can use.
200 Chapter 6 • Plot Statements

STAT=AUTO | SUM | MEAN

specifies the statistic that is to be computed for the VALUE= column when the
column is numeric.
AUTO
computes the SUM statistic when the VALUE= column is numeric. When the
column is character, it uses the first column value as the statistic value.
SUM | MEAN
computes the SUM or MEAN statistic when the VALUE= column is numeric.
When the column is character, it uses the first column value as the statistic value.

Default AUTO

Interaction When the VALUE= column is character, the STAT= option always uses
the first column value as the statistic value. In that case, SUM and
MEAN are ignored.

TEXTGROUP=discrete-attr-var
specifies the discrete attribute variable for a discrete attribute map that maps text
attributes to values for each observation. The discrete attribute variable is defined in
a DISCRETEATTRVAR statement.

Restrictions A discrete attribute variable specification must be a direct reference to

the attribute variable. It cannot be set by a dynamic variable.

The SIZE= specification in the discrete attribute map TEXTATTRS=

option is ignored.

Interaction When this option is specified, the COLORGROUP= option is

ignored .

TITLE="string"
specifies the text for the table title.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default No table title is displayed

Note When an axis table is specified in the prototype of a data-driven layout, if

the table is on the X axis, then the table title appears only in the first
column of each row. If the table is on the Y axis, then the table title appears
only in the first row of each column.

Tip Use the TITLEATTRS= option to control the appearance of the table title.

TITLEATTRS=style-element | style-element(text-options) | (text-options)

specifies the color and font attributes of the table title.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default The GraphLabelText style element.

Interaction If one or more text options are specified and they do not include all the
font properties such as color, family, size, weight, and style, then the
properties that are not specified are derived from the GraphLabelText
style element.
AXISTABLE Statement 201

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element.

“Text Options” on page 1351 for available text-options.

TITLEHALIGN=AUTO | CENTER | LEFT | RIGHT

specifies the horizontal alignment of the axis table header label relative to the axis
table width.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
AUTO
aligns the title based on table type as follows:
• For a Y-axis table, aligns the title according to the effective TITLEJUSTIFY=
option value.
• Starting with the third maintenance release of SAS 9.4, for an X-axis table,
aligns the title LEFT.
CENTER | LEFT | RIGHT
horizontally aligns the table title center, left, or right.

Default AUTO

Restriction In the second maintenance release of SAS 9.4, this option applies only
to Y-axis tables. Starting with the third maintenance release of SAS 9.4,
this option applies to Y-axis tables and to X-axis tables.

Interaction The TITLE= option must be specified for this option to have any effect.

TITLEJUSTIFY=LEFT | CENTER | RIGHT

specifies the justification of the title string. The justification is relative to the axis
table width.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default LEFT

Restriction This option applies only to Y-axis tables.

Interaction The TITLE= option must be specified for this option to have any effect.

VALUEATTRS=style-element | style-element(text-options) | (text-options)

specifies the color and font attributes of the text values.

Default The GraphDataText style element.

Interaction If one or more text options are specified and they do not include all the
font properties such as color, family, size, weight, and style, then the
properties that are not specified are derived from the GraphLabelText
style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

202 Chapter 6 • Plot Statements

VALUEFORMAT=format
specifies a SAS format or a user-defined format for the table values.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default The format that is in effect for the column specified in the VALUE= option.
If no format is in effect, BEST6 is used for numeric columns.

Note Not all of the SAS formats are supported. See Appendix 4, “SAS Formats
Not Supported,” on page 1353.

VALUEHALIGN=AUTO | LEFT | CENTER | RIGHT

in a Y-axis table, specifies the horizontal alignment of the column values relative to
the column width.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
AUTO
uses the effective value of the VALUEJUSTIFY= option.
LEFT | CENTER | RIGHT
aligns the values left, center, or right relative to the column width.

Default AUTO

Restriction This option applies only to Y-axis tables.

VALUEJUSTIFY=AUTO | LEFT | CENTER | RIGHT

specifies the justification of the values in the axis table.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
AUTO
uses LEFT for text values or RIGHT for numeric values.
CENTER | LEFT | RIGHT
horizontally aligns the table values center, left, or right, relative to the column
width.

Default AUTO

Restriction This option applies only to Y-axis tables.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.
Example: AXISTABLE Statement 203

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
The AXISTABLE statement enables you to place text values along the X or Y axis at
specific values on the axis. It offers more flexibility than the BLOCKPLOT statement,
which is used to denote changes in block values along the axis. The X and Y data does
not need to be sorted.

Example: AXISTABLE Statement

This example shows how to add a table of average sales data by division below a bar
chart of total sales by product and country. Here is the output that is generated by this
example.

An inner margin is created at the bottom of the layout container to reserve space for the
table. An AXISTABLE statement is used in the INNERMARGIN block to show the
average sales by division for each product.
Here is the SAS code for this example.
proc template;
define statgraph axistable;
begingraph;
entrytitle "Average Product Sales By Division and Country";
layout overlay / cycleattrs=true
204 Chapter 6 • Plot Statements

yaxisopts=(offsetmax=0.15 label="Sales By Country");

innermargin / align=bottom opaque=true backgroundcolor=cxf5f5f5;
axistable x=product value=actual /
name="division" stat=mean display=(label)
headerlabel="Sales By Division"
headerlabelattrs=GraphLabelText
valueattrs=(size=9pt weight=bold)
colorgroup=division class=division;
endinnermargin;
barchart category=product y=actual / name="country"
barlabel=true barlabelformat=dollar5.0
stat=mean group=country groupdisplay=cluster;
discretelegend "country" / title="Country:" location=inside
valign=top;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.prdsale template=axistable;

run;

See Also
“Creating an Axis-Aligned Inset with an Axis Table” in SAS Graph Template Language:
User's Guide

BANDPLOT Statement
Creates a band plot that typically shows confidence or prediction limits.
Requirements: You must specify either an X= argument or a Y= argument. You cannot specify both..
When you specify the X argument, you must also specify LIMITLOWER and
LIMITUPPER arguments for Y values.
When you specify the Y argument, you must also specify LIMITLOWER and
LIMITUPPER arguments for X values.
The plot data should be sorted by the X or Y variable that is used in the BANDPLOT
statement. Otherwise, specify CONNECTORDER= AXIS in the BANDPLOT
statement.

Syntax
BANDPLOT X=column | expression
LIMITLOWER=number | numeric-column | expression
LIMITUPPER=number | numeric-column | expression </option(s)>;
BANDPLOT Y=numeric-column | expression
LIMITLOWER=number | numeric-column | expression
LIMITUPPER=number | numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
BANDPLOT Statement 205

ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.
CONNECTORDER=VALUES | AXIS
specifies how to connect the data points to form the band lines.
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the band labels.
DATATRANSPARENCY=number
specifies the degree of the transparency of the band fill and band outline.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display an outlined band area, a filled band area, or an
outlined and filled band area.
EXTEND=TRUE | FALSE
specifies whether the constant or "step" band is to be drawn to the area
bounded by the axes.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled band area.
INDEX=positive-integer-column | expression
specifies indices for mapping band attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.
JUSTIFY=LEFT | CENTER | RIGHT
specifies the location of the data point relative to the step when TYPE=STEP.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the band outlines.
TYPE=SERIES | STEP
specifies how the data points for lower and upper band boundaries are
interpolated.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
band plot.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the band labels relative to the plot area.
CURVELABELLOWER="string" | column
206 Chapter 6 • Plot Statements

specifies a label for the lower band limit.

CURVELABELPOSITION=AUTO | MAX | MIN |START | END
specifies the position of the band labels relative to the curve line.
CURVELABELUPPER="string" | column
specifies a label for the upper band limit.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate band plot for each unique group value of the specified
column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

MODELNAME="plot-name"
specifies the name of the plot from which to derive the interpolation for the
band.
NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
You must specify either an X= or Y= argument. You cannot specify both. In addition, the
LIMITLOWER= and LIMITUPPER= arguments must be used to specify the lower and
upper lines for the band.
X=column | expression
specifies X values. Numeric or character values can be used.

Requirement You must also specify the LIMITLOWER= and the LIMITUPPER=
arguments for the Y values.

Y=column | expression
specifies Y values. Numeric or character values can be used.

Requirement You must also specify the LIMITLOWER= and the LIMITUPPER=
arguments for the X values.

LIMITLOWER=number | numeric-column | expression

specifies a constant or column representing the values of the lower band line.

Interactions When this option is used with the X= option, it specifies the Y value
or values.

When this option is used with the Y= option, it represents the X value
or values.

Note If a constant is specified, then a straight line is drawn.

LIMITUPPER=number | numeric-column | expression

specifies a constant or column representing the values of the lower band line.
BANDPLOT Statement 207

Interactions When this option is used with the X= option, it specifies the Y value
or values.

When this option is used with the Y= option, it represents the X value
or values.

Note If a constant is specified, then a straight line is drawn.

Optional Arguments
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
AUTO
specifies that anti-aliasing is controlled by the ANTIALIAS= option in the ODS
GRAPHICS statement.
OFF
specifies that anti-aliasing is always disabled for this plot.

Default AUTO

Interaction This option overrides the ANTIALIAS= option in the ODS

GRAPHICS statement.

CONNECTORDER=VALUES | AXIS
specifies how to connect the data points to form the band lines.
VALUES
connects data points in the order read from the X column (or Y column).
AXIS
connects data points as they occur left-to-right along the X axis (or bottom-to-top
along the Y axis).

Tip You can use this value to ensure the expected connect order for certain
types of series lines (for example, time series) when the input data might
not be sorted by the X column (or Y column).

Default VALUES

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the band labels.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, text color is derived from the

GraphData1:ContrastColor–GraphDataN:ContrastColor style references.
The font is derived from the GraphValueText style element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.
208 Chapter 6 • Plot Statements

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOWER="string" | column
specifies a label for the lower band limit.

Default No curve label is displayed for the lower band

Requirements For non-grouped data, use "string".

For grouped data, use a column to define the lower band labels for
each group value. All of the labels for a specific group value must
be the same. Otherwise, the results are unpredictable.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELUPPER="string" | column
specifies a label for the upper band limit.

Default No curve label is displayed for the upper band

Requirements For non-grouped data, use "string".

For grouped data, use a column to define the upper band labels for
each group value. All of the labels for a specific group value must
be the same. Otherwise, the results are unpredictable.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the band labels relative to the plot area.
INSIDE
locates the labels inside the plot area
OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the BANDPLOT is used in multicell

layouts such as LATTICE, DATAPANEL, or DATALATTICE where
axes might be external to the grid.

Interaction This option is used with the CURVELABELPOSITION= option to

determine where the band labels appear. For more information, see
“Location and Position of Curve Labels” on page 185.

CURVELABELPOSITION=AUTO | MAX | MIN |START | END

specifies the position of the band labels relative to the curve line.
AUTO
positions the band labels automatically near the band boundary along unused
axes whenever possible (typically Y2 and X2).
BANDPLOT Statement 209

Restriction This option is used only when

CURVELABELLOCATION=OUTSIDE.

MAX
forces the band labels to appear near maximum band values (maximum X-values
for horizontal curves, and maximum Y-values for vertical curves).
MIN
forces the band label to appear near minimum band values (minimum X-values
for horizontal curves, and minimum Y-values for vertical curves)
START
forces band labels to appear near the beginning of the curve.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

END
forces band labels to appear near the end of the curve.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION=INSIDE

is specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Interaction This option is used with the CURVELABELLOCATION= option to

determine where the band labels appear. For more information, see
“Location and Position of Curve Labels” on page 185.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN=

in an axis statement, the data points that are used to determine the
position of the band label might fall outside of the graph area. In that
case, the band label might not be displayed or might be positioned
incorrectly.

DATATRANSPARENCY=number
specifies the degree of the transparency of the band fill and band outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the band curve labels.

Tip The FILLATTRS= option can be used to set transparency for just the band
area. You can combine this option with FILLATTRS= to set one
210 Chapter 6 • Plot Statements

transparency for the band outline but a different transparency for the band
fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display an outlined band area, a filled band area, or an outlined
and filled band area.
STANDARD
displays filled band with no outline
ALL
displays an outlined, filled band
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays an outlined band

FILL displays a filled band

Default The value of the DisplayOpts attribute of the GraphBand style element,
which is DisplayOpts="FILL" by default.

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the band.

EXTEND=TRUE | FALSE
specifies whether the constant or "step" band is to be drawn to the area bounded by
the axes.

Default FALSE

Requirement When this option is used for a constant band, constants must be
specified for the upper and lower band limits. This requirement does
not apply to "step" bands.

Interactions This option is ignored when band labels are placed inside the plot
area (CURVELABELLOCATION=INSIDE). To extend the bands in
that case, use the CURVELABELLOCATION=OUTSIDE option.

If the X or Y value is character, then the EXTEND= option is

honored only when the upper and lower limits specify a number.

Tip If this option is not specified, then there can be a small gap between
the line and the axis. The gap is controlled by the axis offset. If the
axis offset is set to 0, then there is no gap.

See “boolean ” on page 1339 for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled band area.

Defaults For non-grouped data, the GraphConfidence:Color style reference.

For grouped data, the Color attribute of GraphData1–GraphDataN style

elements.
BANDPLOT Statement 211

Interaction For this option to have any effect, the fill must be enabled by the ODS
style or by the DISPLAY= option.

Tip The DATATRANSPARENCY= option sets the transparency for both

the band fill and band outline. You can combine this option with
DATATRANSPARENCY= to set one transparency for the band outline
and a different transparency for the band fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element.

“Fill Options” on page 1348 for available fill-options values.

GROUP=column | discrete-attr-var | expression

creates a separate band plot for each unique group value of the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Defaults If the band outline is enabled by the ODS style or the DISPLAY=
option, then each distinct group value is represented in the plot by a
different combination of outline color (defined by the ContrastColor
attribute of the GraphData1–GraphDataN and GraphMissing style
elements) and outline pattern (defined by the LineStyle attribute of the
GraphData1–GraphDataN and GraphMissing style elements).

If the band fill is enabled by the ODS style or the DISPLAY= option,
then each distinct group value is represented in the plot by a different
fill color (defined by the Color attribute of the GraphData1–
GraphDataN and GraphMissing style elements).

Restriction This option can be used only when a numeric column is specified for
the upper limit or the lower limit of the band plot. The other limit
could be a constant, if desired.

Interactions To label grouped band plots, you must specify

CURVELABELLOWER= column and CURVELABELUPPER=
column

The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of colors and line
patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Tip The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the band lines, but the
212 Chapter 6 • Plot Statements

PATTERN= suboption of the OUTLINEATTRS= option could be used

to assign the same line pattern to all band outlines.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping band attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

JUSTIFY=LEFT | CENTER | RIGHT

specifies the location of the data point relative to the step when TYPE=STEP.
BANDPLOT Statement 213

Default LEFT

Requirement TYPE= STEP must also be specified for this option to have any
effect.

Interaction If the MODELNAME= option is specified, then this option is

ignored.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

MODELNAME="plot-name"
specifies the name of the plot from which to derive the interpolation for the band.
When this option is used, the band plot forms prediction or confidence limits for the
plot that supplies the fitted model.

Requirement plot-name must be the name that has been assigned on the associated
plot’s NAME= option.

Interaction This option overrides the JUSTIFY= and TYPE= options.

Tip If this option is not specified, then the interpolation is set by the
TYPE= option.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the band outlines.

Defaults For non-grouped data, the GraphConfidence style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements.
214 Chapter 6 • Plot Statements

Interaction For this option to have any effect, the outline must be enabled by the
ODS style or by the DISPLAY= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element.

“Line Options” on page 1349 for available line-options values.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, LIMITUPPER, LIMITLOWER, GROUP,
CURVELABELUPPER, and CURVELABELLOWER.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the band plot.
If this option is used, then it replaces all of the information that is displayed by
default. Roles for columns that do not contribute to the band plot can be specified
along with roles that do.
(role-list)
an ordered, space-separated list of unique BANDPLOT and user-defined roles.
BANDPLOT roles include X , Y , LIMITUPPER , LIMITLOWER , GROUP ,
INDEX , CURVELABELUPPER , and CURVELABELLOWER .
User-defined roles are defined with the ROLENAME= option.

Note CURVELABELUPPER and CURVELABELLOWER are considered

roles only when they are assigned a column of values. They are not
considered roles and do not display data tips when assigned a string.

Example This example displays data tips for the columns assigned to the roles
X, LIMITUPPER, and LIMITLOWER as well as the column Obs,
which is not assigned to any pre-defined BANDPLOT role. The Obs
column must first be assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X LIMITUPPER LIMITLOWER)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X , Y , LIMITUPPER , LIMITLOWER , and
GROUP .
BANDPLOT Statement 215

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TYPE=SERIES | STEP
specifies how the data points for lower and upper band boundaries are interpolated.
SERIES
connects the data points using line segments (as in a SeriesPlot).
STEP
connects the data points (as in a StepPlot).
216 Chapter 6 • Plot Statements

Default SERIES

Interactions TYPE=STEP must be specified to enable the JUSTIFY= option.

If the MODELNAME= option is specified, then this option is ignored.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interactions This option is ignored if the X= argument is not specified.

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interactions This option is ignored if the Y= argument is not specified.

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
A band plot can specify an X column with Y upper and lower limits, or a Y column with
X upper and lower limits. If you specify the X argument, then you must specify
LIMITLOWER and LIMITUPPER arguments for the Y values to apply the limits to the
Y axis. If you specify the Y argument, then you must specify LIMITLOWER and
LIMITUPPER arguments for the X values to apply the limits to the X axis.
When you use a BANDPLOT statement to display prediction or confidence limits, the
band plot can be used with another plot that specifies a fitted model. For example, it can
be used with a series or step plot. In these cases, use the BANDPLOT option
MODELNAME= or TYPE= to identify the interpolation for the band.
You can use the BANDPLOT statement in displays that are independent of other plots.
For example, a band plot can be used to define yellow and green areas in an OVERLAY
LAYOUT statement that also contains a scatter plot. This use implies concern for any of
the scatter plot values that fall in the yellow area and comfort for any values that fall in
the green area. For this use, the upper and lower limits would be specified by a constant.
Note: The BANDPLOT statement is optimized to work as a Confidence or Prediction
band. If the band is self intersecting (not sorted for X or for Y), then the resulting
band is unpredictable. With unsorted data, the band that is generated for an output
Raster Image might not match the band that is generated for an output Vector
Graphic.
Example: BANDPLOT Statement 217

Example: BANDPLOT Statement

The following graph was generated by the “Example Program” on page 217:

Example Program
Here is the code for this example.
proc template;
define statgraph bandplot;
begingraph;
entrytitle "Fit Plot for Weight";
layout overlay;
bandplot x=height limitupper=uppermean
limitlower=lowermean /
name="band" modelname="fit"
legendlabel="95% Confidence Limits";
scatterplot x=height y=weight / primary=true;
seriesplot x=height y=predict / name="fit"
legendlabel="Fit Line";
discretelegend "fit" "band";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.classfit template=bandplot;

run;
218 Chapter 6 • Plot Statements

BARCHART Statement
Creates a bar chart computed from input data.
Tips: For charts that have a large number of bars that are very close together, slight
variations in spacing that normally occur due to integer rounding can become more
obvious. Subpixel rendering provides more precise bar spacing in that case. In the
second maintenance release of SAS 9.4 and in earlier releases, specify
SUBPIXEL=ON in the BEGINGRAPH statement to enable subpixel rendering. See
SUBPIXEL= on page 33. Starting with the third maintenance release of SAS 9.4,
subpixel rendering is enabled by default.
To disable subpixel rendering in the third maintenance release of SAS 9.4 and in
later releases, specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an
ODS GRAPHICS statement. For information about the BEGINGRAPH statement
SUBPIXEL= option, see SUBPIXEL= on page 33. For information about the ODS
GRAPHICS statement SUBPIXEL= option, see “ODS GRAPHICS Statement” in
SAS ODS Graphics: Procedures Guide.

Syntax
BARCHART CATEGORY=column | expression </option(s)>;
BARCHART CATEGORY=column | expression
RESPONSE=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.
COLORBYFREQ=TRUE | FALSE
specifies whether the bar colors are based on statistical values when the
COLORRESPONSE= option is not specified.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option or the
COLORBYFREQ= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the bar colors to
a continuous color gradient.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bar connect lines.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, connect
line, and bar labels, if displayed.
DISPLAY=STANDARD | ALL | (display-options)
specifies which bar features to display.
BARCHART Statement 219

DISPLAYZEROLENGTHBAR=TRUE | FALSE
specifies whether zero-length bars are drawn.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled bar area.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled bar area.
FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.
INDEX=positive-integer-column | expression
specifies indices for mapping bar attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.
INTERVALBARWIDTH=dimension
specifies the width of the bars in an interval bar chart as a ratio of the interval
width.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bar outlines.
TARGET=numeric-column | expression
specifies the target value for each bar.

Axes options
BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
BARLABEL=TRUE | FALSE
specifies whether the bar statistic value is displayed at the end of the bar.
BARLABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the text properties of the bar label text.
BARLABELFITPOLICY=AUTO | NONE
specifies a policy for avoiding collisions among the bar labels when labels
are displayed.
BARLABELFORMAT=format
220 Chapter 6 • Plot Statements

specifies the text format used to display the bar label.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
SEGMENTLABEL=TRUE | FALSE
specifies whether a label is displayed inside each bar segment.
SEGMENTLABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the text properties of the text for the bar segment label.
SEGMENTLABELFITPOLICY=NONE | NOCLIP | THIN
specifies a policy for fitting the bar segment labels within the bar segments.
SEGMENTLABELFORMAT=format
specifies the text format for the bar segment labels.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all bars from the category midpoints.
GROUP=column | discrete-attr-var | expression
creates a separate bar segment or bar for each unique group value in the
specified column.
GROUP100=NONE | MAGNITUDE | POSITIVE
displays the computed response values (FREQ, SUM, or MEAN), normalized
to 100%.
GROUPDISPLAY=STACK | CLUSTER
specifies how to display grouped bars.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when the bar is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
COLORSTAT=FREQ | PCT | SUM | MEAN | PROPORTION
specifies the statistic to be calculated for the data range of the bar-color
gradient.
STAT=FREQ | PCT | SUM | MEAN | PROPORTION
specifies the statistic to be computed for the Y-axis.

Required Arguments
Specifying only CATEGORY= creates a bar chart with bars that, by default, represent
frequency counts or percents of CATEGORY. Specifying both CATEGORY= and
BARCHART Statement 221

RESPONSE= creates a bar chart with bars representing summarized values of

RESPONSE categorized by CATEGORY.
CATEGORY=column | expression
specifies the column or expression for the category values.

Notes You can use X= as an alternative to CATEGORY=. If you use X=, then be
aware that the TIP=, TIPFORMAT=, and TIPLABEL= options recognize X
as the category role and not as CATEGORY.

For interval category values, if a user-defined format is applied to the

category column, the format should map each category value to only one
unique formatted value. Otherwise, unexpected results might occur.

RESPONSE=numeric-column | expression
specifies the numeric column or expression for the response values.

Notes You can use Y= as an alternative to RESPONSE=. If you use Y=, then be
aware that the TIP=, TIPFORMAT=, and TIPLABEL= options will
recognize Y as the response role and not RESPONSE in that case.

This option is required only when you want summarized values of

RESPONSE that are categorized by CATEGORY.

Optional Arguments
BARLABEL=TRUE | FALSE
specifies whether the bar statistic value is displayed at the end of the bar. For
grouped clustered bars, each bar is labeled with the summarized value of the bar. For
grouped stacked bars, the segmented bar is labeled with the accumulated,
summarized value of all the bar segments.

Default FALSE

Tip The font and color attributes for the label are specified by the
BARLABELATTRS= option. The text format is specified by the
BARLABELFORMAT= option.

See “boolean ” on page 1339 for other Boolean values that you can use.

BARLABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the bar label text.

Default The GraphDataText style element.

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

BARLABELFITPOLICY=AUTO | NONE
specifies a policy for avoiding collisions among the bar labels when labels are
displayed.
222 Chapter 6 • Plot Statements

AUTO
for a vertical bar chart, rotates the bar labels if the labels exceed the midpoint
spacing. For a horizontal bar chart, always draws the labels horizontally. The
following figure shows an example.

See the BARWIDTH= option for more information about the bar spacing.
NONE
does not rotate the bar labels. Labels that are too long overlap.
Labels can collide along their length and along their height. In some cases, if one or
more labels collide when the specified fit policy is used, then all of the labels are
dropped from the display. When that occurs, the following warning message is
written to the SAS log:
WARNING: The bar labels are suppressed. Use BARLABELFITPOLICY=NONE to
force the labels to be displayed.

TIP If the labels collide along their height, then using the BARLABELATTRS=
option to reduce the label font size might eliminate the collision.

Default AUTO

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

BARLABELFORMAT=format
specifies the text format used to display the bar label.

Default The column format assigned to the RESPONSE= column or BEST6

if no format is assigned.

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.

Default 0.85

Range 0.1–1, where 0.1 is the narrowest and 1 is the widest

BARCHART Statement 223

Interaction Starting with the third maintenance release of SAS 9.4, the
INTERVALBARWIDTH= option overrides this option for an interval
bar chart.

Notes This option is needed only to change the default behavior.

By default, the bar width automatically adjusts based on the number of

bars to be displayed and the wall width.

Tip To remove any inter-bar gap, set BARWIDTH=1.

BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.

Default The GraphAxisLines style element.

Notes The baseline is always drawn by default.

When style-element is specified, only the style element’s COLOR,

LINESTYLE, and LINETHICKNESS attributes are used.

Tip To suppress the baseline, set the line thickness to 0:

baselineattrs=(thickness=0)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline. The baseline is always
displayed in the chart, whether for a specified value or for the default value. When
this option is used, the axis range is adjusted to include the baseline, and the baseline
is placed at the specified value on the response axis.

Default 0

Interactions If GROUPDISPLAY=STACKED is specified, then this option is

ignored and the baseline is not displayed.

This option is ignored when the GROUP100= option is used.

If necessary, the response axis data range is extended to include the

baseline intercept. When a logarithmic response axis is requested and
BASELINEINTERCEPT= specifies 0 or a negative value, the
224 Chapter 6 • Plot Statements

response axis reverts to a linear axis. To restore the log axis in that
case, set BASELINEINTERCEPT= to a positive value.

Note Label positions are automatically adjusted to prevent the labels from
overlapping.

Tips Control the appearance of the baseline with the BASELINEATTRS=

option.

To suppress the baseline, use the BASELINEATTRS= option to set

the line thickness to 0.

The baseline does not add a tick or a tick value to the axis. To label the
baseline, use a REFERENCELINE statement to overlay a line with the
same X or Y value and include the CURVELABEL= option to specify
the label text.

COLORBYFREQ=TRUE | FALSE
specifies whether the bar colors are based on statistical values when the
COLORRESPONSE= option is not specified. Setting this option to TRUE enables
you to color the bars based on frequency counts, percentages, or proportions.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default FALSE

Requirement The COLORSTAT= option must be FREQ, PCT, or PROPORTION

for this option to have any effect.

Interactions This option is ignored when the COLORRESPONSE= option is

specified.

When the GROUP= option is specified with the COLORBYFREQ=

option, the color attributes are controlled by the COLORBYFREQ=
option.

The COLOR= suboption of the FILLATTRS=,

FILLPATTERNATTRS=, and OUTLINEATTRS= options overrides
this option for the associated color attribute.

Note This option is independent of the STAT= and RESPONSE= options.

Tips Use the COLORSTAT= option to specify whether frequency counts,

percentages, or proportions are computed for the
COLORRESPONSE= column.

Use the FILLTYPE= option to specify whether each bar is filled with
a solid color or with a gradient color.

See “Example 3: Bar Chart with Bar Colors Controlled by a Statistic” on

page 249

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option or the
COLORBYFREQ= option.
BARCHART Statement 225

color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Defaults For outline-only bars, the ThreeColorAltRamp style element

For bars with fill, the ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option or the
COLORBYFREQ=TRUE option must also be specified.

Tip Use the DISPLAY= option to specify whether outlines and fills are
displayed.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the bar colors to a
continuous color gradient.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. Each bar is colored using one color from the gradient
range. When a range attribute map variable is specified, the colors that are defined in
the associated range attribute map are used instead.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column has
multiple values for a single GROUP value, unexpected results might
occur.
226 Chapter 6 • Plot Statements

Interactions The COLORBYFREQ= option is ignored when this option is

specified.

When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

When fill, fill pattern, or both are displayed, this option overrides
suboption COLOR= in the FILLATTRS= option and in the
FILLPATTERNATTRS= option and varies the color according to the
color gradient or the attribute map.

When only the outlines are displayed, this option overrides suboption
COLOR= in the OUTLINEATTRS= option and varies the outline
color according to the color gradient or the attribute map.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

Use the COLORSTAT= option to specify the statistic to compute for

the COLORRESPONSE= column.

Use the FILLTYPE= option to specify whether each bar is filled with
a solid color or with a gradient color.

For a numeric column or expression, the ThreeColorRamp style

element defines the fill color gradient, and the ThreeColorAltRamp
style element defines the outline color gradient.

COLORSTAT=FREQ | PCT | SUM | MEAN | PROPORTION

specifies the statistic to be calculated for the data range of the bar-color gradient.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
The statistics that are available depend on the COLORRESPONSE= and
COLORBYFREQ= option specifications. When the COLORRESPONSE= option is
specified, the following values are valid:

SUM
MEAN
When the COLORRESPONSE= option is not specified and
COLORBYFREQ=TRUE is in effect, the following values are valid:

FREQ frequency count

PCT percentages between 0 and 100
PROPORTION proportions between 0 and 1

Defaults FREQ when the COLORRESPONSE= option is not specified and

COLORBYFREQ=TRUE is in effect.

SUM when the COLORRESPONSE= option is specified.

Interactions This option is ignored when the COLORRESPONSE= option is not

specified and COLORBYFREQ=FALSE is in effect.
BARCHART Statement 227

This option might affect existing SAS programs. For programs written
before the third maintenance release of SAS 9.4, if STAT= and
COLORRESPONSE= are specified in a BARCHART statement, then
the bar-chart colors and color statistic might change from those of the
previous SAS releases. To restore the original colors and color statistic
in that case, set COLORSTAT= in the BARCHART statement to the
same statistic that is specified in STAT=.

Note This option is independent of the STAT= and RESPONSE= options.

See COLORBYFREQ= on page 224

COLORRESPONSE= on page 225

STAT= on page 240

“Example 3: Bar Chart with Bar Colors Controlled by a Statistic” on

page 249

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar connect lines.

Default The GraphConnectLine style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows bars
with each of the skins applied.

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
228 Chapter 6 • Plot Statements

the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

When FILLTYPE=GRADIENT is in effect, DATASKIN=SHEEN is

ignored. In that case, use one of the other skins.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, connect line, and
bar labels, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the filled
bar area. You can combine this option with FILLATTRS= to set one
transparency for the bar outlines and connect lines but a different
transparency for the bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEOFFSET=number
specifies an amount to offset all bars from the category midpoints.

Default 0 (no offset, all bars are centered on the category midpoints)

Range -0.5 to +0.5, where 0.5 represents half the distance between category ticks.
Normally, a positive offset is to the right when ORIENT=VERTICAL, and
up when ORIENT=HORIZONTAL. (If the layout's axis options set
REVERSE=TRUE, then the offset direction is also reversed.)

Tip Setting the discrete offset for the plots does not affect the axis minimum
and maximum offsets. In some cases, setting a discrete offset can cause
clipping at each end of the axis. In those cases, use the OFFSETMIN= and
OFFSETMAX= axis options to increase the axis minimum and maximum
offsets to accommodate the discrete offset.

See “About the DISCRETEOFFSET= Option” on page 245

Chapter 8, “Axis Options in Layouts,” on page 889 for information about

the REVERSE=, OFFSETMIN=, and OFFSETMAX= axis options

ORIENT=

DISPLAY=STANDARD | ALL | (display-options)

specifies which bar features to display.
BARCHART Statement 229

STANDARD
displays outlined, filled bars
ALL
displays outlined, filled bars, and connect lines
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:
OUTLINE
displays outlined bars.
FILL
displays bars with a solid fill.
CONNECT
displays line segments connecting adjacent midpoints at the end of each bar.
FILLPATTERN
displays bars with a patterned fill. This setting is used primarily for grouped
bar charts that must be rendered in monochrome for use in a journal article.
The fill patterns make it easier to distinguish among groups when color is not
available.

Default STANDARD

Restriction Connect lines are not drawn for grouped data.

Note The connect lines are drawn in axis order starting with the third
maintenance release of SAS 9.4. They are drawn in data order in prior
releases.

Tips Use the OUTLINEATTRS=, FILLATTRS=, and

FILLPATTERNATTRS= options to control the appearance of the bars.

Use CONNECTATTRS= to control the appearance of the connect lines.

You can specify both FILL and FILLPATTERN to combine solid fills
and pattern fills in the bars.

DISPLAYZEROLENGTHBAR=TRUE | FALSE
specifies whether zero-length bars are drawn.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
A zero-length bar is displayed as a line spanning the normal bar width at the bar-
chart baseline on the response axis. When this option is set to TRUE, zero-length
bars are displayed. Otherwise, they are suppressed. The following figure shows a
simple example of each outcome. In the figure, the plot wall outline, category axis
line, and bar-chart baseline are suppressed for clarity.
230 Chapter 6 • Plot Statements

Default TRUE

Interaction This option is ignored when the GROUP= and

GROUPDISPLAY=STACK options are in effect. In that case, zero-
length bar segments are drawn.

Note When this option is set to FALSE, the bar is not drawn, but other
elements associated with the bar such as the target bar, the error bar, the
bar label, and the data label, are drawn.

Tip This option is useful when the bar-chart baseline is suppressed.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled bar area.

Defaults For non-grouped data, the GraphDataDefault:Color style reference

For grouped data, the Color attribute of the GraphData1–GraphDataN

style elements.

Interaction When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is
ignored, and the bar fill colors vary according to the gradient.

Tip The DATATRANSPARENCY= option sets the transparency for the bar
fills, bar outlines, and connect lines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the bar outlines
and connect lines but a different transparency for the bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled bar area.
style-element
specifies the name of a style element. You can specify only one of the elements
GraphData1–GraphDataN.

Restriction The only styles that are delivered by SAS that support fill patterns
are JOURNAL2, JOURNAL3, and MONOCHROMEPRINTER. If
any other such style is in effect and this option uses style-element in
its specification, then this option is ignored.

(fill-pattern-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:
COLOR=color | style-reference
specifies a color to use for the bar-fill-pattern lines. With grouped data, the
COLOR= setting has the effect of holding the fill color constant across all
group values.
PATTERN=line-pattern
specifies a line pattern to use for the bar fill.
BARCHART Statement 231

To specify a line-pattern, combine a line-direction prefix (R for right, L for

left, and X for cross hatch) with a line-identification number:

With grouped data, the PATTERN= setting has the effect of holding the fill
pattern constant across all group values.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

See DISPLAY=

FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
SOLID
each bar is filled with the color that is assigned to that bar.
GRADIENT
an alpha gradient is used to determine the bar fill color. Each bar is filled with a
color and a transparency gradient that starts at the bar top with the specified fill
color and transparency, and transitions to fully transparent at the bar baseline.
The initial fill color is determined by a style element or by the FILLATTRS=
option COLOR= suboption. The initial transparency is determined by the
DATATRANSPARENCY= option or by the FILLATTRS= option
TRANSPARENCY= suboption.

Interactions The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

In the second maintenance release of SAS 9.4,

FILLTYPE=GRADIENT is ignored when
GROUPDISPLAY=STACK is in effect. Starting with the third
maintenance release of SAS 9.4, FILLTYPE=GRADIENT is
honored in that case.

Tips Use the DATATRANSPARENCY= option or the FILLATTRS=

option TRANSPARENCY= suboption to set the initial
transparency in the gradients.

For grouped plots, use the FILLATTRS= option in a discrete

attribute map to set the initial transparency in the gradients for
specific values.
232 Chapter 6 • Plot Statements

See DATASKIN= on page 227

Default SOLID

Interaction The DISPLAY= option must include FILL for this option to have any
effect.

GROUP=column | discrete-attr-var | expression

creates a separate bar segment or bar for each unique group value in the specified
column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

For example, the sashelp.Cars data used in the “Example Program” on page 247
contains a column named Origin, which identifies the region that produces each car.
This column could be used in the BARCHART statement to group the bars in the
display (see the GROUPDISPLAY= option to see the output for the grouped bars):
layout overlay;
barchart category=type response=mpg_highway /
stat=mean group=origin name="b";
discretelegend "b" / title="Regions:";
endlayout;

Defaults If bar fills or fill patterns are enabled by the ODS style or by the
DISPLAY= option, then each distinct group value is represented in the
plot by a different fill color or fill pattern. The fill colors are defined
by the Color attribute of the GraphData1–GraphDataN and
GraphMissing style elements. The fill patterns are defined by the
FillPattern attribute of the GraphData1–GraphDataN and
GraphMissing style elements.

If bar outlines are enabled by the ODS style or by the DISPLAY=

option, then each distinct group value is represented in the plot by a
different outline. The outline colors are defined by the ContrastColor
attribute of the GraphData1–GraphDataN and GraphMissing style
elements.

Interactions Connect lines are not drawn for grouped data.

By default, the group values are mapped in the order of the data. Use
the GROUPORDER= option to control the sorting order of the
grouped bar segments. Use the INDEX= option to alter the default
sequence of colors and line patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

When both the GROUP= and COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.
BARCHART Statement 233

Note The bar display depends on the setting for the GROUPDISPLAY=
option.

Tip The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the bar outlines, but you can
use the PATTERN= setting on the OUTLINEATTRS= option to assign
the same line pattern to all bar outlines and connect lines.

See “DISCRETEATTRVAR Statement” on page 1297

GROUP100=NONE | MAGNITUDE | POSITIVE

displays the computed response values (FREQ, SUM, or MEAN), normalized to
100%.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
NONE
displays the summarized data.
MAGNITUDE
normalizes both the negative and positive values to 100% by magnitude, and
displays the group values, preserving the sign. The positive values are displayed
above the bars for a vertical bar chart and on the right end for a horizontal bar
chart. The negative values are displayed enclosed in parentheses below the bars
for a vertical bar chart and on the left end for a horizontal bar chart.
The following figure illustrates the effect of MAGNITUDE on stacked bars in a
vertical bar chart.

POSITIVE
drops the negative values and normalizes only the positive values to 100%. The
following figure demonstrates the effect of POSITIVE on clustered bars in a
vertical bar chart. This chart uses the same data as the chart in the previous
figure.
234 Chapter 6 • Plot Statements

Notice that the negative values are dropped from the chart.

Default NONE

Requirement The GROUP= option must be specified for this option to have any
effect.

Interaction When this option is used, the BASELINEINTERCEPT= and

TARGET= options are ignored.

Note You can use this option with any value for the GROUPDISPLAY=
option.

Tip To display the values, specify BARLABEL=TRUE.

GROUPDISPLAY=STACK | CLUSTER
specifies how to display grouped bars.
STACK
displays group values as stacked segments within the category bar.

CLUSTER
displays group values as separate adjacent bars that replace the single category
bar. Each cluster of group values is centered at the category midpoint on the axis.
This example illustrates the clusters and also how groups are displayed when
they have an unequal number of unique values.
BARCHART Statement 235

Default STACK

Interaction When you use the BARLABEL= option and the GROUP= option, the
BARLABEL values are displayed for each bar when
GROUPDISPLAY=CLUSTER. When GROUPDISPLAY=STACK, the
whole bar is labeled at the top.

Tip For a linear response axis, when STAT=MEAN or STAT=PCT, the axis
tick values might be displayed as integer values when
GROUPDISPLAY=STACK. Changing GROUPDISPLAY= to
CLUSTER in that case might cause the axis values to change to
decimal values. To keep the integer axis values in both cases, you can
specify the INTEGER=TRUE option for the response axis. See
INTEGER= on page 915.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
DESCENDING
orders the groups within a category in descending order.

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.
236 Chapter 6 • Plot Statements

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping bar attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.
BARCHART Statement 237

INTERVALBARWIDTH=dimension
specifies the width of the bars in an interval bar chart as a ratio of the interval width.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default The width specified by the BARWIDTH= option.

Restriction This option applies only to a linear or time category axis. When the
category axis is discrete, this option is ignored.

Interaction When the category data is interval, this option overrides the
BARWIDTH= option.

Tip To make the category axis type linear or time, include TYPE=LINEAR
or TYPE=TIME in the category axis options or assign the role of
primary plot to a plot that makes the category axis linear or time.

See “dimension” on page 1340

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The response-variable label. If a label is not defined, then the response-
variable name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.

Default VERTICAL

Notes When this option is set to HORIZONTAL, the category variable appears on
the Y (or Y2) axis and the response variable appears on the X (or X2) axis.
To set the axis properties for this chart, you should use the appropriate axis
options of the layout container.

When this option is set to VERTICAL, the category variable appears on the
X (or X2) axis and the response variable appears on the Y (or Y2) axis. To
set the axis properties for this chart, you should use the appropriate axis
options of the layout container.
238 Chapter 6 • Plot Statements

If you change the orientation of the bar chart, then you should adjust the
layout container’s axis options appropriately.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar outlines.

Defaults For non-grouped data, the ContrastColor, LineThickness, and

LineStyle attributes of the GraphOutlines style element.

For grouped data and filled bars, the ContrastColor attribute of the
GraphData1–GraphDataN style elements, and the LineThickness and
LineStyle attributes of the GraphOutlines style element.

For grouped data and unfilled bars, the ContrastColor and LineStyle
attributes of the GraphData1–GraphDataN style elements, and the
LineThickness attribute of the GraphOutlines style element.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the bar outline colors vary according to the gradient.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

SEGMENTLABEL=TRUE | FALSE
specifies whether a label is displayed inside each bar segment.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
BARCHART Statement 239

For an ungrouped bar chart or for a grouped bar chart with

GROUPDISPLAY=CLUSTER, AUTO displays a bar label inside each bar. The label
displays the statistic for that bar. For a grouped bar chart with
GROUPDISPLAY=STACK, AUTO displays a label inside each bar segment. Each
segment label displays the statistic for that bar segment, as shown in the following
figure.

When this value is set to FALSE, no labels are displayed inside the bars.

Default FALSE

Tips For a grouped bar chart with GROUPDISPLAY=STACK, specify both

SEGMENTLABEL=TRUE and BARLABEL=TRUE to display a label for
each bar segment and a label for the entire bar.

Use the SEGMENTLABELATTRS= option to modify the appearance of

the label text.

Use the SEGMENTLABELFITPOLICY= option to specify a policy for

fitting the labels inside the bars.

Use the SEGMENTLABELFORMAT= option to modify the format of the

segment labels.

See “boolean ” on page 1339 for other Boolean values that you can use.

SEGMENTLABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the text for the bar segment label.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default The GraphDataText style element.

Interaction This option is ignored when SEGMENTLABEL=FALSE.

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element.

“Text Options” on page 1351 for available text-options.

240 Chapter 6 • Plot Statements

SEGMENTLABELFITPOLICY=NONE | NOCLIP | THIN

specifies a policy for fitting the bar segment labels within the bar segments.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
NONE
no attempt is made to fit each segment label within its bar. Long bar segment
labels might overlap other graphical elements. The segment labels are not
considered when the axis ranges are computed. As a result, segment labels that
extend beyond the plot area are clipped.
NOCLIP
does not clip bar segment labels that extend beyond the plot area. Labels that do
not fit within the plot area extend into the graph axis area and might overlap axis
elements.
THIN
drops any bar segment label that does not fit within its segment. For a vertical bar
chart, the label width must not exceed the bar width, and the text height must not
exceed the segment height. For a horizontal bar chart, the label text height must
not exceed the bar width, and the label length must not exceed the segment
length.

Default THIN

Interaction This option is ignored when SEGMENTLABEL=FALSE.

SEGMENTLABELFORMAT=format
specifies the text format for the bar segment labels.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default The column format assigned to the RESPONSE= column, or BEST6 if

no format is assigned.

Interaction This option is ignored when SEGMENTLABEL=FALSE.

STAT=FREQ | PCT | SUM | MEAN | PROPORTION

specifies the statistic to be computed for the Y-axis. For bar charts with no
RESPONSE= column:

FREQ frequency count

PCT percentages between 0 and 100
PROPORTION proportions between 0 and 1
Note: Prior to SAS 9.4, PCT displayed proportions between 0 and 1. To restore the
original PCT results in SAS 9.4 and later releases, specify PROPORTION
instead.
For bar charts with a RESPONSE= column:

SUM
MEAN

Defaults For bar charts with no RESPONSE= column, the default is FREQ.

For bar charts with a RESPONSE= column, the default is SUM.

BARCHART Statement 241

Note When this option is used with the GROUP=group option, the specified
statistic is computed for each segment that is created for the unique group
values.

Tip If this option is used with COLORRESPONSE= in SAS programs that

were written before the third maintenance release of SAS 9.4, the bar-
chart colors and color statistic might change from those of the previous
SAS releases. To restore the original colors and color statistic, set
COLORSTAT= in the BARCHART statement to the same statistic that is
specified in STAT=.

TARGET=numeric-column | expression
specifies the target value for each bar. The visual representation is a triangle with a
line at the target value.
layout overlay;
barchart category=type response=mpg_highway / barwidth=.8
target=mpg_city group=origin groupdisplay=cluster
name='bar';
discretelegend 'bar';
endlayout;

Default No targets are displayed.

Interactions For this option to take effect, the RESPONSE= argument must also be
used.

If the GROUP= option is used and GROUPDISPLAY= STACK, then

this option is ignored.

This option is ignored when the GROUP100= option is used.

Tips The statistic indicated by the STAT= option applies to the

TARGET=column. If a constant value is desired for each target, then
specify it only once for repeated category (X) values (or category and
GROUP combinations), and leave the other target values missing.

The target color is that of the bar outline.

242 Chapter 6 • Plot Statements

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar. If this
option is used, then the information specified replaces all of the information that is
displayed by default.
(role-list)
an ordered, space-separated list of unique BARCHART roles. BARCHART roles
include CATEGORY or X, RESPONSE or Y, COLORRESPONSE, INDEX,
GROUP, and TARGET.

Notes For the category and response roles, the TIP= option recognizes only
the category and response arguments that you use in the BARCHART
statement. If you use the CATEGORY= and RESPONSE= arguments,
then you must specify roles CATEGORY and RESPONSE.
Conversely, if you use the X= and Y= arguments, then you must
specify roles X and Y.

The COLORRESPONSE role is valid starting with the third

maintenance release of SAS 9.4.

Example The following example displays data tips for the columns assigned
only to the roles CATEGORY and RESPONSE:

TIP=(CATEGORY RESPONSE)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: CATEGORY or X, RESPONSE or Y,
COLORRESPONSE, and GROUP.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement with the IMAGEMAP option specified,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIP=(RESPONSE)
TIPFORMAT=(RESPONSE=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.
BARCHART Statement 243

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example TIP=(RESPONSE)
TIPLABEL=(RESPONSE="Average Sales")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

URL=string-column
specifies an HTML page to display when the bar is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
bar that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable bars, you must include an ODS GRAPHICS

ON statement that has the IMAGEMAP option specified, and you
must write the output to the ODS HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Notes For non-grouped data, the values of the column are expected to be
same for each unique category value. If they are not, then the results
might be unpredictable.

For grouped data, the values of the column are expected to be the
same for each unique category and GROUP combination.

Tips The URL value can be blank for some category values, meaning that
no action is taken when the bars for those category values are
selected.

The URL value can be the same for different category values,
meaning that the same action is taken when the bars for those
category values are selected.
244 Chapter 6 • Plot Statements

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interactions This option is ignored if the RESPONSE= argument is not specified.

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details

Statement Description
The BARCHART statement creates a bar chart with bars that represent summarized
response values. The response values are categorized by the unique category values or,
starting with the third maintenance release of SAS 9.4, by the bins in binned category
data. The BARCHART statement takes nonsummarized data as input and calculates the
appropriate summarization statistics (sum, mean, and so on) for each unique category
value or category bin. Prior to the third maintenance release of SAS 9.4, the category
axis for a bar chart must be discrete. Starting with the third maintenance release of SAS
9.4, the category axis can be discrete, linear, or time. The response axis in all cases is
interval.
When the chart is oriented vertically, the X (or X2) axis is used for CATEGORY and the
Y (or Y2) axis is used for RESPONSE. When it is oriented horizontally, the X (or X2)
axis is used for RESPONSE and the Y (or Y2) axis is used for CATEGORY. (See
ORIENT= on page 237.)
By default, if the category column is character, then the bars in the chart appear in the
order in which the category values are present in the input data. If the category column is
numeric, then the values are presented in ascending order. For non-grouped data,
duplicated category values are summarized into a unique value. For grouped data, the
category values are summarized as needed. (See the GROUP= option.)
Starting with the third maintenance release of SAS 9.4, for numeric category values, an
interval bar chart is generated only when the category axis type is linear or time. To
specify a category axis type of linear or time, include the TYPE= option in the category
axis options, or assign the role of primary plot to a plot that sets the category axis type to
linear or time automatically. By default, a bar is drawn for each unique category value,
which can result in a large number of bars for numeric category data.
When binning is used, for each bin, a summarization statistic is computed, and a bar is
drawn that represents that statistic. The width of each bar spans the width of the bin that
it represents. The left-most edge of the bar represents the start of the bin, and the right-
most edge represents the end. See “Example 1: Horizontal Bar Chart” on page 247.
BARCHART Statement 245

TIP Prior to the third maintenance release of SAS 9.4, use the HISTOGRAM
statement to create a bar chart that represents response values along an interval axis.

About the DISCRETEOFFSET= Option

The DISCRETEOFFSET= option is useful for graphing multiple response variables side
by side on a common axis. By default within an overlay-type layout, if multiple
BARCHART statements are used with different response variables, then the bars for
matching category values are centered on the midpoints and the bars are superimposed.
To make it easier to distinguish among superimposed bars, you can assign a different
BARWIDTH= setting to each BARCHART statement in the overlay:
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchart category=year response=A_revenue / stat=sum name="A"

legendlabel="A" barwidth=0.8 ;
barchart category=year response=B_revenue / stat=sum name="B"
legendlabel="B" barwidth=0.6 ;
barchart category=year response=C_revenue / stat=sum name="C"
legendlabel="C" barwidth=0.4 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;

To place the different response values side by side, you can assign a different offset to
each BARCHART statement. The BARWIDTH= option can be used with
DISCRETEOFFSET= to create narrower bars that require less width within the plot
area:
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchart category=year response=A_revenue / stat=sum name="A"

legendlabel="A"
discreteoffset=-0.3 barwidth=0.3 ;
barchart category=year response=B_revenue / stat=sum name="B"
legendlabel="B"
246 Chapter 6 • Plot Statements

discreteoffset=0 barwidth=0.3 ;
barchart category=year response=C_revenue / stat=sum name="C"
legendlabel="C"
discreteoffset=+0.3 barwidth=0.3 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;

Different combinations of DISCRETEOFFSET and BARWIDTH can be used to get the

effect that you want. Gaps can be created between bars by providing a narrower bar
width. Or, bars can be overlapped if the bar widths are increased in proportion to the
discrete offset.
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchart category=year response=A_revenue / stat=sum name="A"

legendlabel="A" datatransparency=0.2
discreteoffset=-0.2 barwidth=0.5 ;
barchart category=year response=B_revenue / stat=sum name="B"
legendlabel="B" datatransparency=0.2
discreteoffset=0 barwidth=0.5 ;
barchart category=year response=C_revenue / stat=sum name="C"
legendlabel="C" datatransparency=0.2
discreteoffset=+0.2 barwidth=0.5 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;
Example 1: Horizontal Bar Chart 247

Examples

Example 1: Horizontal Bar Chart

The following graph was generated by the “Example Program” on page 247:

Example Program
proc template;
define statgraph barchart;
begingraph;
entrytitle "Average Mileage by Vehicle Type";
layout overlay;
barchart category=type response=mpg_highway /
stat=mean orient=horizontal;
endlayout;
248 Chapter 6 • Plot Statements

endgraph;
end;

proc sgrender data=sashelp.cars template=barchart;

run;

Example 2: Interval Bar Chart

Interval bar charts are available starting with the third maintenance release of SAS 9.4.
In the second maintenance release of SAS 9.4 and in earlier releases, use the
HISTOGRAM statement to generate an interval bar chart. The following graph was
generated by the “Example Program” on page 248:

Example Program
Here is the SAS code.
proc template;
define statgraph cylinders;
begingraph;
entrytitle "Interval Bar Chart of Vehicle Engine Cylinders";
layout overlay /
xaxisopts=(label="Engine Cylinders" type=linear
linearopts=(tickvaluelist=(3 4 5 6 8 10 12)))
yaxisopts=(label="Percentage of Vehicles Manufactured"
griddisplay=on linearopts=(tickvalueformat=percent7.1));
barchart category=cylinders / stat=proportion
barlabel=true barlabelformat=percent7.1;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=cylinders;

Example 3: Bar Chart with Bar Colors Controlled by a Statistic 249

run;

Details
An interval bar chart can be generated only when the category axis type is LINEAR or
TIME. In this example, the TYPE=LINEAR option is included in the XAXISOPTS=
options. With numeric category data, a bar is drawn for each unique category value. In
some cases, that can generate too many bars in the resulting chart. In this example, there
are only seven unique values. The TICKVALUELIST= option is used in the
XAXISOPTS= option to display all of the values on the category axis.

Example 3: Bar Chart with Bar Colors Controlled by a Statistic

The ability to use a computed statistic to control the bar colors in a bar chart is available
starting with the third maintenance release of SAS 9.4. This example uses the
COLORBYFREQ=TRUE option to enable a computed statistic to control the bar colors
and the COLOSTAT=PCT to specify percentage as the controlling statistic. Here is the
output from “Example Program” on page 249.

Example Program
proc template;
define statgraph barchart;
begingraph;
entrytitle "Average Mileage by Vehicle Type";
layout overlay;
barchart category=type response=mpg_highway / name="bar"
stat=mean orient=horizontal
colorbyfreq=true colorstat=pct;
continuouslegend "bar" /
title="Percent of Total Models Manufactured";
endlayout;
endgraph;
end;
run;
250 Chapter 6 • Plot Statements

proc sgrender data=sashelp.cars template=barchart;

run;

BARCHARTPARM Statement
Creates a bar chart specified by pre-summarized data.
Requirement: The input data must be pre-summarized, with appropriate summarization statistics
(sum, mean, and so on) computed for the RESPONSE column.
Tips: For charts that have a large number of bars that are very close together, slight
variations in spacing that normally occur due to integer rounding can become more
obvious. Subpixel rendering provides more precise bar spacing in that case. In the
second maintenance release of SAS 9.4 and in earlier releases, specify
SUBPIXEL=ON in the BEGINGRAPH statement to enable subpixel rendering. See
SUBPIXEL= on page 33. Starting with the third maintenance release of SAS 9.4,
subpixel rendering is enabled by default.
To disable subpixel rendering in the third maintenance release of SAS 9.4 and in
later releases, specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an
ODS GRAPHICS statement. For information about the BEGINGRAPH statement
SUBPIXEL= option, see SUBPIXEL= on page 33. For information about the ODS
GRAPHICS statement SUBPIXEL= option, see “ODS GRAPHICS Statement” in
SAS ODS Graphics: Procedures Guide.

Syntax
BARCHARTPARM CATEGORY=column | expression
RESPONSE=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing
or bin width.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the bar colors to
a continuous color gradient.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bar connect lines.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, error bars,
connect line, and data labels, if displayed.
BARCHARTPARM Statement 251

DISPLAY=STANDARD | ALL | (display-options)

specifies which bar features to display.
DISPLAYZEROLENGTHBAR=TRUE | FALSE
specifies whether zero-length bars are drawn.
ERRORBARATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the error bars associated with the bars.
ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
ERRORLOWER=numeric-column | expression
specifies the values of the lower endpoints on the Y error bars.
ERRORUPPER=numeric-column | expression
specifies the values of the upper endpoints on the Y error bars.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled bar area.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled bar area.
FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.
INDEX=positive-integer-column | expression
specifies indices for mapping bar attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.
INTERVALBARWIDTH=dimension
specifies the width of the bars in an interval bar chart as a ratio of the interval
width.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bar outlines.
TARGET=numeric-column | expression
specifies the target value for each bar.

Axes options
BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar.
TIPFORMAT=(role-format-list)
252 Chapter 6 • Plot Statements

specifies display formats for tip columns.

TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
specifies the label that appears at the end of each bar.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the labels that are specified in the
DATALABEL= option.
DATALABELFITPOLICY=AUTO | NONE | ROTATE | SPLIT | SPLITALWAYS
specifies a policy for avoiding collisions among the bar labels when labels
are displayed.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed data
labels.
DATALABELTYPE=AUTO | COLUMN
specifies whether the data labels display the RESPONSE values or the values
of the column that is specified by the DATALABEL= option.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
SEGMENTLABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the text properties of the text for the bar segment label.
SEGMENTLABELFITPOLICY=NONE | NOCLIP | THIN
specifies a policy for fitting the bar segment labels within the bar segments.
SEGMENTLABELFORMAT=format
specifies the text format for the bar segment labels.
SEGMENTLABELTYPE=NONE | AUTO
specifies whether a label is displayed inside each bar segment.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all bars from the category midpoints.
GROUP=column | discrete-attr-var | expression
creates a separate bar segment or bar for each unique group value in the
specified column.
GROUP100=NONE | MAGNITUDE | POSITIVE
displays the response values, normalized to 100%.
GROUPDISPLAY=STACK | CLUSTER
specifies how to display grouped bars.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
BARCHARTPARM Statement 253

specifies an HTML page to display when the bar is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
CATEGORY=column | expression
specifies the column for the unique category values. All values are treated as
discrete.

Note You can use X= as an alternative to CATEGORY=. If you use X=, then be
aware that the TIP=, TIPFORMAT=, and TIPLABEL= options will recognize
X as the category role and not CATEGORY in that case.

RESPONSE=numeric-column | expression
specifies the column for the response values.

Note You can use Y= as an alternative to RESPONSE=. If you use Y=, then be
aware that the TIP=, TIPFORMAT=, and TIPLABEL= options will recognize
Y as the response role and not RESPONSE in that case.

Optional Arguments
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.

Default 0.85

Range 0.1–1, where 0.1 is the narrowest and 1 is the widest

Interaction Starting with the third maintenance release of SAS 9.4, the
INTERVALBARWIDTH= option overrides this option for an interval
bar chart.

Notes This option is needed only to change the default behavior.

By default, the bar width automatically adjusts based on the number of

bars to be displayed and the wall width.

Tip To remove any inter-bar gap, set BARWIDTH=1.

BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.

Default The GraphAxisLines style element.

Notes The baseline is always drawn by default.

When style-element is specified, only the style element’s COLOR,

LINESTYLE, and LINETHICKNESS attributes are used.

Tip To suppress the baseline, set the line thickness to 0:

baselineattrs=(thickness=0)
254 Chapter 6 • Plot Statements

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline. The baseline is always
displayed in the chart, whether for a specified value or for the default value. When
this option is used, the axis range is adjusted to include the baseline, and the baseline
is placed at the specified value on the response axis.

Default 0

Interactions If GROUPDISPLAY=STACKED is specified, then this option is

ignored and the baseline is not displayed.

If necessary, the response axis data range is extended to include the

baseline intercept. When a logarithmic response axis is requested and
BASELINEINTERCEPT= specifies 0 or a negative value, the
response axis reverts to a linear axis. To restore the log axis in that
case, set BASELINEINTERCEPT= to a positive value.

Note Label positions are automatically adjusted to prevent the labels from
overlapping.

Tips Control the appearance of the baseline with the BASELINEATTRS=

option.

To suppress the baseline, use the BASELINEATTRS= option to set

the line thickness to 0.

The baseline does not add a tick or a tick value to the axis. To label the
baseline, use a REFERENCELINE statement to overlay a line with the
same X or Y value and include the CURVELABEL= option to specify
the label text.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or bin
width.
BARCHARTPARM Statement 255

Default 0.85

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.

Requirement For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

Interaction When GROUPDISPLAY=CLUSTER, the default BARWIDTH is

1.0.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Defaults For outline-only bars, the ThreeColorAltRamp style element

256 Chapter 6 • Plot Statements

For bars with fill, the ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip Use the DISPLAY= option to specify whether outlines and fills are
displayed.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the bar colors to a
continuous color gradient.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. Each bar is colored using one color from the gradient
range. When a range attribute map variable is specified, the colors that are defined in
the associated range attribute map are used instead.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column has
multiple values for a single GROUP value, unexpected results might
occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

When fill, fill pattern, or both are displayed, this option overrides
suboption COLOR= in the FILLATTRS= option and in the
FILLPATTERNATTRS= option and varies the color according to the
color gradient or the attribute map.

When only the outlines are displayed, this option overrides suboption
COLOR= in the OUTLINEATTRS= option and varies the outline
color according to the color gradient or the attribute map.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

Use the FILLTYPE= option to specify whether each bar is filled with
a solid color or with a gradient color.

For a numeric column or expression, the ThreeColorRamp style

element defines the fill color gradient, and the ThreeColorAltRamp
style element defines the outline color gradient.
BARCHARTPARM Statement 257

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar connect lines.

Default The GraphConnectLine style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

DATALABEL=column | expression
specifies the label that appears at the end of each bar.

Interactions Starting with the second maintenance release of SAS 9.4, this option is
ignored when DATALABELTYPE=AUTO.

When the GROUP= option is in effect, the data label values are
displayed only when GROUPDISPLAY=CLUSTER.

If the GROUP= option is in effect and there are multiple input

observations per bar for the GROUP= column, then the value for the
DATALABEL= column should be the same for each observation that
is on the same bar.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the labels that are specified in the
DATALABEL= option.

Default The GraphDataText style element.

Interaction For this option to take effect, the DATALABEL= option must also be
used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

DATALABELFITPOLICY=AUTO | NONE | ROTATE | SPLIT | SPLITALWAYS

specifies a policy for avoiding collisions among the bar labels when labels are
displayed.
AUTO
selects a collision avoidance policy based on the chart orientation and data type.
For a numeric column with ORIENT=VERTICAL, AUTO rotates the labels if
they do not fit the midpoint spacing. For a character column, AUTO splits the
labels if they do not fit the midpoint spacing.

Note When ORIENT=HORIZONTAL, AUTO always draw the labels

horizontally.

Tip If character labels do not fit after splitting, then try using ROTATE instead
of AUTO.

See ORIENT= on page 273 for information about chart orientation.

BARWIDTH= for information about bar spacing.

258 Chapter 6 • Plot Statements

NONE
does not attempt to fit bar labels that collide.
ROTATE
rotates the bar labels for vertical bars if the labels collide in the available width.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

SPLIT
splits the label for vertical bars at a split character only if a split is needed at that
character in order to make the label fit the available space. No split occurs at split
characters that occur where a split is not needed. If the label does not contain any
of the specified split characters, then a split does not occur. In that case, if the
label does not fit the available space, then it might collide with the adjoining
labels.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

See the DATALABELSPLITCHAR= option for information about

specifying the split characters

SPLITALWAYS
splits the label for vertical bars at every occurrence of a split character. If the
label does not contain any of the specified split characters, then a split does not
occur.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

See the DATALABELSPLITCHAR= option for information about

specifying the split characters

Here is an example of a vertical bar chart where DATALABELFITPOLICY=AUTO

and a numeric column is used as the data labels.

In this case, AUTO rotates the numeric labels to avoid collision.

In some cases, if one or more labels collide when the specified fit policy is used, then
all of the labels are dropped from the display. When that occurs, the following
warning message is written to the SAS log:
WARNING: The bar labels are suppressed. Use DATALABELFITPOLICY=NONE to
BARCHARTPARM Statement 259

force the labels to be displayed.

Default AUTO

Requirement The DATALABEL= option must also be specified.

Interaction When DATALABELTYPE=AUTO is in effect, for a vertical bar

chart, only AUTO, NONE, and ROTATE are valid. All other values
revert to AUTO. For a horizontal bar chart, only AUTO and NONE
are valid. All other values revert to AUTO.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split. When multiple
split characters are specified, each character in the list is treated as a separate split
character unless the specified characters appear consecutively in the data label. In
that case, all of the specified split characters together are treated as a single split
character.
When DATALABELFITPOLICY=SPLIT and a data label collision is detected, the
data label is split on a specified split character only if a split is needed at that point in
order to make the label fit. In that case, a split might not occur on every split
character. When DATALABELFITPOLICY=SPLITALWAYS, the data label is split
unconditionally on every occurrence of a split character. If the data label does not
contain any of the specified split characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

The DATALABELFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interactions The DATALABELFITPOLICY= option specifies the policy that is

used to manage the split behavior of the data label.

The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the displayed data label or are
dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed data labels.
260 Chapter 6 • Plot Statements

TRUE
drops a split character from the display when a split occurs at that character. Split
characters at which a split does not occur are left in place. The
DATALABELFITPOLICY= option determines where the labels are split. When
DATALABELFITPOLICY=SPLIT, each label is split at a split character only
where a split is needed in order to make the label fit the available space. At each
split point, the split character is dropped, and the characters that follow the split
character, up to but not including the split character at the next split point, are
wrapped to the following line.
When DATALABELFITPOLICY=SPLITALWAYS, each label is split at every
instance of a split character. All of the split characters are dropped. The
characters that follow each split character, up to but not including the next split
character, are wrapped to the next line.
The following figure shows how label Product*Group*1 is split when the
DATALABELSPLITCHARDROP=TRUE and DATALABELSPLITCHAR="*"
options are specified with the SPLIT and SPLITALWAYS fit policies.

In this example, when DATALABELFITPOLICY=SPLIT, the label is split at the

first occurrence of the asterisk in order to make the label fit. No split is needed at
the second asterisk. The first asterisk is dropped, and Group*1 wraps to the next
line. Notice that the second asterisk is not dropped in this case. When
DATALABELFITPOLICY=SPLITALWAYS, the label is split at every
occurrence of the asterisk. In this case, both asterisks are dropped, and the
characters that follow each asterisk wrap to the next line.
FALSE
includes the split characters in the data label display. The
DATALABELFITPOLICY= option determines how the split characters are
displayed. When DATALABELFITPOLICY=SPLIT, each data label is split at a
split character only where a split is needed in order to make the label fit the
available space. A split might not occur at every split character in the label. At
each split point, the split character remains as the last character in the current
line. The characters that follow the split character, up to and including the split
character at the next split point, are then wrapped to the following line. This
process repeats until the entire data label is displayed.
When DATALABELFITPOLICY=SPLITALWAYS, each data label is split at
every instance of a split character in the label regardless of whether a split is
actually needed. Each split character remains as the last character in the current
line. The characters that follow each split character, up to and including the next
split character, are then wrapped to the next line.
The following figure shows how label Product*Group*1 is split when the
DATALABELSPLITCHARDROP=FALSE and DATALABELSPLITCHAR="*"
options are specified with the SPLIT and SPLITALWAYS fit policies.
BARCHARTPARM Statement 261

In this example, when DATALABELFITPOLICY=SPLIT, the label is split at the

first occurrence of the asterisk in order to make the label fit. No split is needed at
the second asterisk. The characters that follow the first asterisk wrap to the next
line. When DATALABELFITPOLICY=SPLITALWAYS, the label is split at
every occurrence of the asterisk. Each asterisk remains as the last character in the
current line, and the characters that follow are wrapped to the next line.

Default TRUE. A split character is dropped from the data-label display

when a split occurs at that character.

Requirements The DATALABEL= option must also be specified.

The DATALABELFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELTYPE=AUTO | COLUMN
specifies whether the data labels display the RESPONSE values or the values of the
column that is specified by the DATALABEL= option.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
AUTO
the labels are displayed as follows:
• For an ungrouped bar chart, the summarized value for each bar is placed
above the bar.
• For a grouped bar chart with stacked bars, the total of the summarized
segment values for each bar is placed above the segmented bar.
• For a grouped bar chart with clustered bars, the summarized value for each
bar in the cluster is placed above the bar.

Interactions AUTO overrides the DATALABEL= option.

When AUTO is in effect, some data-label fit policies are

unavailable. See DATALABELFITPOLICY=.

COLUMN
the data labels display the DATALABEL= column values.

Interaction The DATALABEL= option must be specified for COLUMN to

have any effect. If the DATALABEL= option is not specified,
AUTO is used instead.
262 Chapter 6 • Plot Statements

Default COLUMN

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows bars
with each of the skins applied.

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

When FILLTYPE=GRADIENT is in effect, DATASKIN=SHEEN is

ignored. In that case, use one of the other skins.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, error bars, connect
line, and data labels, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

BARCHARTPARM Statement 263

Tip The FILLATTRS= option can be used to set transparency for just the bar
fills. You can combine this option with FILLATTRS= to set one
transparency for the bar outlines, error bars, and connect lines but a
different transparency for the bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEOFFSET=number
specifies an amount to offset all bars from the category midpoints.

Default 0 (no offset, all bars are centered on the category midpoints)

Range -0.5 to +0.5, where 0.5 represents half the distance between category ticks.
Normally, a positive offset is to the right when ORIENT=VERTICAL, and
up when ORIENT=HORIZONTAL. (If the layout's axis options set
REVERSE=TRUE, then the offset direction is also reversed.)

Tip Setting the discrete offset for the plots does not affect the axis minimum
and maximum offsets. In some cases, setting a discrete offset can cause
clipping at each end of the axis. In those cases, use the OFFSETMIN= and
OFFSETMAX= axis options to increase the axis minimum and maximum
offsets to accommodate the discrete offset.

See “About the DISCRETEOFFSET= Option” on page 280

Chapter 8, “Axis Options in Layouts,” on page 889 for information about

the OFFSETMIN= and OFFSETMAX= axis options

ORIENT=

DISPLAY=STANDARD | ALL | (display-options)

specifies which bar features to display.
STANDARD
displays outlined, filled bars
ALL
displays outlined, filled bars and also connect lines
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:
OUTLINE
displays outlined bars.
FILL
displays bars with a solid fill.
CONNECT
displays line segments connecting adjacent midpoints at the end of each bar.
FILLPATTERN
displays bars with a patterned fill. This setting is used primarily for grouped
bar charts that must be rendered in monochrome for use in a journal article.
The fill patterns make it easier to distinguish among groups when color is not
available.

Default STANDARD
264 Chapter 6 • Plot Statements

Restriction Neither error bars nor connect lines are displayed for grouped data.

Interaction Connect lines are not drawn for grouped data.

Note Error bars are automatically displayed whenever the ERRORUPPER=

or ERRORLOWER= options are specified.

Tips Use the OUTLINEATTRS= , FILLATTRS= , and

FILLPATTERNATTRS= options to control the appearance of the bars.
Use CONNECTATTRS= to control the appearance of the connect lines.

Both FILL and FILLPATTERN can be specified to combine solid fills

and pattern fills in the bars.

DISPLAYZEROLENGTHBAR=TRUE | FALSE
specifies whether zero-length bars are drawn.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
A zero-length bar is displayed as a line spanning the normal bar width at the bar-
chart baseline on the response axis. When this option is set to TRUE, zero-length
bars are displayed. Otherwise, they are suppressed. The following figure shows a
simple example of each outcome. In the figure, the plot wall outline, category axis
line, and bar-chart baseline are suppressed for clarity.

Default TRUE

Interaction This option is ignored when the GROUP= and

GROUPDISPLAY=STACK options are in effect. In that case, zero-
length bar segments are drawn.

Note When this option is set to FALSE, the bar is not drawn, but other
elements associated with the bar such as the target bar, the error bar, the
bar label, and the data label, are drawn.

Tip This option is useful when the bar-chart baseline is suppressed.

ERRORBARATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the error bars associated with the bars.

Defaults For non-grouped data, the GraphError style element.

For grouped data, the LineStyle and LineThickness attributes of the

GraphError style element and the ContrastColor attribute of the
GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.
BARCHARTPARM Statement 265

“Line Options” on page 1349 for available line-options.

ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.

Defaults SERIF in the first maintenance release of SAS 9.4 and earlier releases.

Starting with the second maintanance release of SAS 9.4,

GraphError:CapStyle style reference. If attribute CapStyle is not defined
in the active style, then SERIF is the default value.

Tip The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

ERRORLOWER=numeric-column | expression
specifies the values of the lower endpoints on the Y error bars.

Default The lower segment of the error bars is not drawn.

Requirement The error bar values must be absolute data values, not data values
relative to the value of the bar.

Interaction If the GROUP= option is specified with GROUPDISPLAY=STACK

or with GROUP100=POSITIVE or MAGNITUDE, then this option
is ignored.

Tip You can use the ERRORBARATTRS= option to control the

appearance of the error bars.

ERRORUPPER=numeric-column | expression
specifies the values of the upper endpoints on the Y error bars.

Default The upper segment of the error bars is not drawn.

Requirement The error bar values must be absolute data values, not data values
relative to the value of the bar.

Interaction If the GROUP= option is specified with GROUPDISPLAY=STACK

or with GROUP100=POSITIVE or MAGNITUDE, then this option
is ignored.

Tip You can use the ERRORBARATTRS= option to control the

appearance of the error bars.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled bar area.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

266 Chapter 6 • Plot Statements

For grouped data, the GraphData1:Color–GraphDataN:Color style

references.

Interaction When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is
ignored, and the bar fill colors vary according to the gradient.

Tip The DATATRANSPARENCY= option sets the transparency for bar

fills, bar outlines, error bars, and connect lines. You can combine this
option with DATATRANSPARENCY= to set one transparency for the
bar outlines, error bars, and connect lines but a different transparency
for the bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element .

“Fill Options” on page 1348 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled bar area.
style-element
specifies the name of a style element. You can specify only one of the elements
GraphData1–GraphDataN.

Restriction The only styles that are delivered by SAS that support fill patterns
are JOURNAL2, JOURNAL3, and MONOCHROMEPRINTER. If
any other such style is in effect and this option uses style-element in
its specification, then this option is ignored.

(fill-pattern-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:
COLOR=color | style-reference
specifies a color to use for the bar-fill-pattern lines. With grouped data, the
COLOR= setting has the effect of holding the fill color constant across all
group values.
PATTERN=line-pattern
specifies a line pattern to use for the bar fill.
To specify a line-pattern, combine a line-direction prefix (R for right, L for
left, and X for cross hatch) with a line-identification number:
BARCHARTPARM Statement 267

With grouped data, the PATTERN= setting has the effect of holding the fill
pattern constant across all group values.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

See DISPLAY=

FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
SOLID
each bar is filled with the color that is assigned to that bar.
GRADIENT
an alpha gradient is used to determine the bar fill color. Each bar is filled with a
color and a transparency gradient that starts at the bar top with the specified fill
color and transparency, and transitions to fully transparent at the bar baseline.
The initial fill color is determined by a style element or by the FILLATTRS=
option COLOR= suboption. The initial transparency is determined by the
DATATRANSPARENCY= option or by the FILLATTRS= option
TRANSPARENCY= suboption.

Interactions The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

In the second maintenance release of SAS 9.4,

FILLTYPE=GRADIENT is ignored when
GROUPDISPLAY=STACK is in effect. Starting with the third
maintenance release of SAS 9.4, FILLTYPE=GRADIENT is
honored in that case.

Tips Use the DATATRANSPARENCY= option or the FILLATTRS=

option TRANSPARENCY= suboption to set the initial
transparency in the gradients.

For grouped plots, use the FILLATTRS= option in a discrete

attribute map to set the initial transparency in the gradients for
specific values.

See DATASKIN= on page 262

Default SOLID

Interaction The DISPLAY= option must include FILL for this option to have any
effect.

GROUP=column | discrete-attr-var | expression

creates a separate bar segment or bar for each unique group value in the specified
column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.
268 Chapter 6 • Plot Statements

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

The bar display depends on the setting for the GROUPDISPLAY= option. For
example, for a vertical bar chart with GROUPDISPLAY=STACK, bar segments are
stacked to form the bar. The height of each segment represents the corresponding
group value’s proportional contribution to the response value.

Defaults If bar fills are enabled by the ODS style or by the DISPLAY= option,
then each distinct group value is represented in the plot by a different
fill color or fill pattern. The fill colors are defined by the Color
attribute of the GraphData1–GraphDataN and GraphMissing style
elements. The fill patterns are defined by the FillPattern attribute of
the GraphData1–GraphDataN and GraphMissing style elements.

If bar outlines are enabled by the ODS style or by the DISPLAY=

option, then each distinct group value is represented in the plot by a
different outline. The outline colors are defined by the ContrastColor
attribute of the GraphData1–GraphDataN and GraphMissing style
elements.

Interactions Connect lines are not drawn for grouped data.

By default, the group values are mapped in the order of the data. Use
the GROUPORDER= option to control the sorting order of the group
values.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

When both the GROUP= and COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.

Tips The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the bar outlines, but you
could use the PATTERN= setting on the OUTLINEATTRS= option to
assign the same line pattern to all bar outlines and connect lines.

Use the INDEX= option to alter the default sequence of colors, fill
patterns, and line patterns.

GROUP100=NONE | MAGNITUDE | POSITIVE

displays the response values, normalized to 100%.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
NONE
displays the data as received.
MAGNITUDE
normalizes both the negative and positive values to 100% by magnitude, and
displays the group values, preserving the sign. The positive values are displayed
above the bars for a vertical bar chart and on the right end for a horizontal bar
BARCHARTPARM Statement 269

chart. The negative values are displayed enclosed in parentheses below the bars
for a vertical bar chart and on the left end for a horizontal bar chart.
The following figure illustrates the effect of MAGNITUDE on stacked bars in a
vertical bar chart.

POSITIVE
drops the negative values and normalizes only the positive values to 100%. The
following figure demonstrates the effect of POSITIVE on clustered bars in a
vertical bar chart. This chart uses the same data as the chart in the previous
figure.

Notice that the negative values are dropped from the chart.

Default NONE

Requirement The GROUP= option must be specified for this option to have any
effect.

Interaction Error bars are not drawn when GROUP100=POSITIVE or

MAGNITUDE. See ERRORLOWER= and ERRORUPPER=.

Note You can use this option with any value for the GROUPDISPLAY=
option.
270 Chapter 6 • Plot Statements

Tip To display the values, specify DATALABELTYPE=AUTO.

GROUPDISPLAY=STACK | CLUSTER
specifies how to display grouped bars.
STACK
displays group values as stacked segments within the category bar.

CLUSTER
displays group values as separate adjacent bars that replace the single category
bar. Each cluster of group values is centered at the category midpoint on the axis.
This example illustrates the clusters and also how groups are displayed when
they have an unequal number of unique values.

Default STACK

Interactions When you use the DATALABEL= option and the GROUP= option,
the DATALABEL values are displayed for each bar when
GROUPDISPLAY=CLUSTER. When GROUPDISPLAY=STACK,
the whole bar is labeled at the top.

Error bars are not drawn when GROUPDISPLAY=STACK.

When the TARGET= and GROUP= options are in effect, the target
values are not displayed when GROUPDISPLAY=STACK. In that
BARCHARTPARM Statement 271

case, you must specify GROUPDISPLAY=CLUSTER to display the

target values.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
DESCENDING
orders the groups within a category in descending order.

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.
272 Chapter 6 • Plot Statements

INDEX=positive-integer-column | expression
specifies indices for mapping bar attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

INTERVALBARWIDTH=dimension
specifies the width of the bars in an interval bar chart as a ratio of the interval width.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default The width specified by the BARWIDTH= option.

Restriction This option applies only to a linear or time category axis. When the
category axis is discrete, this option is ignored.

Interaction When the category data is interval, this option overrides the
BARWIDTH= option.

Tip To make the category axis type linear or time, include TYPE=LINEAR
or TYPE=TIME in the category axis options or assign the role of
primary plot to a plot that makes the category axis linear or time.

See “dimension” on page 1340

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The response-variable label. If a label is not defined, then the response-
variable name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.
BARCHARTPARM Statement 273

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.

Default VERTICAL

Notes When this option is set to VERTICAL, the category variable appears on the
X (or X2) axis and the response variable appears on the Y (or Y2) axis. To
set the axis properties for this chart, you should use the appropriate axis
options of the layout container.

When this option is set to HORIZONTAL, the category variable appears on

the Y (or Y2) axis and the response variable appears on the X (or X2) axis.
To set the axis properties for this chart, you should use the appropriate axis
options of the layout container.

If you change the orientation of the bar chart, then you should adjust the
layout container’s axis options appropriately.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar outlines.

Defaults For non-grouped data, the ContrastColor, LineThickness, and

LineStyle attributes of the GraphOutlines style element.

For grouped data and filled bars, the ContrastColor attribute of the
GraphData1–GraphDataN style elements, and the LineThickness and
LineStyle attributes of the GraphOutlines style element.

For grouped data and unfilled bars, the ContrastColor and LineStyle
attributes of the GraphData1–GraphDataN style elements, and the
LineThickness attribute of the GraphOutlines style element.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the bar outline colors vary according to the gradient.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.
274 Chapter 6 • Plot Statements

“Line Options” on page 1349 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles CATEGORY or X, RESPONSE or Y,
ERRORLOWER, ERRORUPPER, GROUP, and INDEX.

SEGMENTLABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the text for the bar segment label.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default The GraphDataText style element.

Interaction This option is ignored when SEGMENTLABELTYPE=NONE.

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element.

“Text Options” on page 1351 for available text-options.

SEGMENTLABELFITPOLICY=NONE | NOCLIP | THIN

specifies a policy for fitting the bar segment labels within the bar segments.
BARCHARTPARM Statement 275

Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
NONE
no attempt is made to fit each segment label within its bar. Long bar segment
labels might overlap other graphical elements. The segment labels are not
considered when the axis ranges are computed. As a result, segment labels that
extend beyond the plot area are clipped.
NOCLIP
does not clip bar segment labels that extend beyond the plot area. Labels that do
not fit within the plot area extend into the graph axis area and might overlap axis
elements.
THIN
drops any bar segment label that does not fit within its segment. For a vertical bar
chart, the label width must not exceed the bar width, and the text height must not
exceed the segment height. For a horizontal bar chart, the label text height must
not exceed the bar width, and the label length must not exceed the segment
length.

Default THIN

Interaction This option is ignored when SEGMENTLABELTYPE=NONE.

SEGMENTLABELFORMAT=format
specifies the text format for the bar segment labels.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Defaults If DATALABELTYPE=AUTO is in effect, the column format assigned

to the RESPONSE= column, or BEST6 if no format is assigned.

If DATALABELTYPE=COLUMN is in effect, the column format

assigned to the DATALABEL= column, or BEST6 if no format is
assigned.

Interaction This option is ignored when SEGMENTLABELTYPE=NONE.

SEGMENTLABELTYPE=NONE | AUTO
specifies whether a label is displayed inside each bar segment.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
For an ungrouped bar chart or for a grouped bar chart with
GROUPDISPLAY=CLUSTER, AUTO displays a bar label inside each bar. The label
for each bar displays the value for that bar. For a grouped bar chart with
GROUPDISPLAY=STACK, AUTO displays a label inside each bar segment. The
label for each bar segment displays the value for that segment.
276 Chapter 6 • Plot Statements

When this value is set to NONE, no labels are displayed inside the bars.

Default NONE

Interaction The DATALABELTYPE= option determines whether the segment

labels display the RESPONSE column values or the values of the
column that is specified by the DATALABEL= option.

Tips For a grouped bar chart with GROUPDISPLAY=STACK, specify both

SEGMENTLABELTYPE=AUTO and DATALABEL=TRUE to display
a label for each bar segment and a label for the entire bar.

Use the SEGMENTLABELATTRS= option to modify the appearance

of the label text.

Use the SEGMENTLABELFITPOLICY= option to specify a policy for

fitting the labels inside the bars.

Use the SEGMENTLABELFORMAT= option to modify the format of

the segment labels.

See “boolean ” on page 1339 for other Boolean values that you can use.

TARGET=numeric-column | expression
specifies the target value for each bar. The visual representation is a triangle with a
line at the target value.
layout overlay;
barchartparm category=type response=mpg_highway / barwidth=.8
target=mpg_city group=origin groupdisplay=cluster
name='bar';
discretelegend 'bar';
endlayout;
BARCHARTPARM Statement 277

Default No targets are displayed.

Interactions For this option to take effect, the RESPONSE= argument must also be
used.

If the GROUP= option is used and GROUPDISPLAY= STACK, then

this option is ignored.

Tip The target color is that of the bar outline.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar. If this
option is used, then the information specified replaces all of the information that is
displayed by default. You can specify roles for columns that do not contribute to the
bar chart along with roles that do.
(role-list)
an ordered, space-separated list of unique BARCHARTPARM roles and user-
defined roles. BARCHARTPARM roles include CATEGORY or X, RESPONSE
or Y, COLORRESPONSE, ERRORUPPER, ERRORLOWER, INDEX,
GROUP, and DATALABEL.

Notes For the category and response roles, the TIP= option recognizes only the
category and response arguments that you use in the BARCHARTPARM
statement. If you use the CATEGORY= and RESPONSE= arguments,
then you must specify roles CATEGORY and RESPONSE. Conversely,
if you use the X= and Y= arguments, then you must specify roles X and
Y.

The COLORRESPONSE role is valid starting with the third

maintenance release of SAS 9.4.

Tip Use the ROLENAME= option to define user-defined roles.

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: CATEGORY or X, RESPONSE or Y,
COLORRESPONSE, ERRORLOWER, ERRORUPPER, and
GROUP.
278 Chapter 6 • Plot Statements

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement with the IMAGEMAP option specified,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

Example The following example displays data tips for the columns assigned to
the roles CATEGORY and RESPONSE as well as for the column Pct.
The Pct column is not assigned to any pre-defined bar chart role, so it
must first be assigned a role.

ROLENAME=(TIP1=PCT)
TIP=(TIP1 CATEGORY RESPONSE)

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)
BARCHARTPARM Statement 279

URL=string-column
specifies an HTML page to display when the bar is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
bar that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirements To generate selectable bars, you must include an ODS GRAPHICS

ON statement that has the IMAGEMAP option specified, and you
must write the output to the ODS HTML destination.

For non-grouped data, the values of the column are expected to be

same for each unique category value. If they are not, then only the
first URL value for a given category value is used.

For grouped data, the values of the column are expected to be the
same for each unique category and GROUP combination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some category values, meaning that
no action is taken when the bars for those category values are
selected.

The URL value can be the same for different category values,
meaning that the same action is taken when the bars for those
category values are selected.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.
280 Chapter 6 • Plot Statements

Details

Statement Description
The input data for the BARCHARTPARM statement is expected to be pre-summarized,
with appropriate summarization statistics (sum, mean, and so on) computed for the
response column. When the category values are not unique, the display is not
meaningful. Prior to the third maintenance release of SAS 9.4, the category axis must be
discrete. Starting with the third maintenance release of SAS 9.4, the category axis can be
discrete, linear, or time. The response axis in all cases is interval.
When the chart is oriented vertically, the X (or X2) axis is used for category and the Y
(or Y2) axis is used for response. When the chart is oriented horizontally, the X (or X2)
axis is used for response and the Y (or Y2) axis is used for category. (See ORIENT= on
page 273.) If the chart is the primary chart, then any charts or plots that are overlaid with
it must have similar axis types.
By default, if the CATEGORY= column is character, then the bars in the chart appear in
the order in which the category values are present in the input data. If the CATEGORY=
column is numeric, then the values are presented in ascending order.

About the DISCRETEOFFSET= Option

The DISCRETEOFFSET= option is useful for graphing multiple response variables side
by side on a common axis. By default within an overlay-type layout, if multiple
BARCHART statements are used with different response variables, then the bars for
matching category values are centered on the midpoints and the bars are superimposed.
To make it easier to distinguish among superimposed bars, you can assign a different
BARWIDTH= setting to each BARCHARTPARM statement in the overlay:
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchartparm category=year response=A_revenue / name="A"

legendlabel="A" barwidth=0.8 ;
barchartparm category=year response=B_revenue / name="B"
legendlabel="B" barwidth=0.6 ;
barchartparm category=year response=C_revenue / name="C"
legendlabel="C" barwidth=0.4 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;
BARCHARTPARM Statement 281

To place the different response values side by side, you can assign a different offset to
each BARCHARTPARM statement. The BARWIDTH= option can be used with
DISCRETEOFFSET= to create narrower bars that require less width within the plot
area:
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchartparm category=year response=A_revenue / name="A"

legendlabel="A"
discreteoffset=-0.3 barwidth=0.3 ;
barchartparm category=year response=B_revenue / name="B"
legendlabel="B"
discreteoffset=0 barwidth=0.3 ;
barchartparm category=year response=C_revenue / name="C"
legendlabel="C"
discreteoffset=+0.3 barwidth=0.3 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;
282 Chapter 6 • Plot Statements

Different combinations of DISCRETEOFFSET and BARWIDTH can be used to get the

effect that you want. Gaps can be created between bars by providing a narrower bar
width. Or, bars can be overlapped if the bar widths are increased in proportion to the
discrete offset.
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchartparm category=year response=A_revenue / name="A"

legendlabel="A" datatransparency=0.2
discreteoffset=-0.2 barwidth=0.5 ;
barchartparm category=year response=B_revenue / name="B"
legendlabel="B" datatransparency=0.2
discreteoffset=0 barwidth=0.5 ;
barchartparm category=year response=C_revenue / name="C"
legendlabel="C" datatransparency=0.2
discreteoffset=+0.2 barwidth=0.5 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;
Example: BARCHARTPARM Statement 283

Example: BARCHARTPARM Statement

The following graph was generated by the “Example Program” on page 283:

Example Program
proc template;
define statgraph barchartparm;
begingraph;
entrytitle "Average Mileage by Vehicle Type";
entryfootnote halign=left
"Error bars show +/- 1 Standard Error";
layout overlay;
barchartparm category=type response=mean /
errorlower=eval(mean-stderr)
errorupper=eval(mean+stderr) ;
endlayout;
endgraph;
end;
run;

/* create summarized data for barchartparm */

proc summary data=sashelp.cars nway;
class type;
var mpg_highway;
output out=mileage mean=mean stderr=stderr ;
run;
284 Chapter 6 • Plot Statements

proc sgrender data=mileage template=barchartparm;

run;

BIHISTOGRAM3DPARM Statement
Creates a three-dimensional bivariate histogram of three variables X, Y, and Z, where the values of X and
Y have been gridded. The Z variable represents a response value for the frequency, percentage counts, or
densities of each bin combination.
Restriction: BIHISTOGRAM3DPARM does not support the data tips that are enabled by the
IMAGEMAP= option in the ODS GRAPHICS statement.
Requirements: The input data must be binned by both X and Y. That is, the values for X column and
Y column must form a complete rectangular grid of bins. Input data with non-binned
columns should be preprocessed with the KDE procedure (SAS/STAT), which
enables you to set the number of bins for X and Y, or with a technique similar to that
used in “Example: BIHISTOGRAM3DPARM Statement” on page 288.
The BIHISTOGRAM3DPARM statement must be specified within a LAYOUT
OVERLAY3D statement and cannot be nested under an OVERLAY,
OVERLAYEQUATED, or PROTOTYPE layout.
The input data for Z= column must be nonnegative.
Note: In the plot display, the direction of the Z axis is upward rather than outward.

Syntax
BIHISTOGRAM3DPARM X=numeric-column | expression
Y=numeric-column | expression
Z=non-negative-numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the bins.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display outlined bins, filled bins, or outlined and filled
bins.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled bins.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bin outlines.
XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input X values represent midpoints, lower endpoints, or
upper endpoints of the bins.
YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input Y values represent midpoints, lower endpoints, or
upper endpoints of the bins.

Axes options
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.
BIHISTOGRAM3DPARM Statement 285

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.

Label options
ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin
endpoints (TRUE) or at the bin midpoints (FALSE).
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the bin location of the numeric X values.
Y=numeric-column | expression
specifies the bin location of the numeric Y values.
Z=nonnegative-numeric-column | expression
specifies the response values, such as the frequency counts, percentages, or densities.

Optional Arguments
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.
TRUE
specifies that the ENDLABELS= option determines how the axis ticks and value
labels are displayed.
FALSE
specifies that standard axes are used. Bin boundaries and midpoints that are set
by the ENDLABELS= option are ignored.

Default TRUE

Interactions For this option to take effect, this plot must be the primary plot in the
parent OVERLAY3D layout. For more information, see the
PRIMARY= option.

When this option is set to TRUE, some X-axis options that are set on
the parent layout might not apply, such as INTEGER=,
TICKVALUELIST=, TICKVALUESEQUENCE=, and
INCLUDERANGES=.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bins.

Default 0
286 Chapter 6 • Plot Statements

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display outlined bins, filled bins, or outlined and filled bins.
STANDARD
does the following based on the layout in which it is contained:
• When used inside an overlay layout, it displays an outlined, filled block
without text values or a label (OUTLINE FILL).
• When used as a stand-alone plot inside a lattice or gridded layout, it displays
an outlined, filled block with text values and a label (OUTLINE FILL
VALUES LABEL).
ALL
displays outlined, filled bins.
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays outlined bins

FILL displays filled bins

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the bins.

ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin endpoints
(TRUE) or at the bin midpoints (FALSE).

Default FALSE.

Interactions This option is ignored if this plot is not the primary plot in the parent
layout. For more information, see the PRIMARY= option.

This option is ignored if BINAXIS= FALSE. By default,

BINAXIS=TRUE.

See “boolean ” on page 1339 for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled bins.

Default The GraphDataDefault style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

BIHISTOGRAM3DPARM Statement 287

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bin outlines.

Default The GraphOutlines style element.

Restriction This option uses only the color specification in the style element or line
options. The line pattern and line thickness specifications are ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

COLOR= on page 1349

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input X values represent midpoints, lower endpoints, or upper
endpoints of the bins.

Default MIDPOINTS

YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input Y values represent midpoints, lower endpoints, or upper
endpoints of the bins.

Default MIDPOINTS

Details
Because the BIHISTOGRAM3DPARM statement does not perform a binning
computation on the input columns, you must pre-bin the data. In addition, input data for
288 Chapter 6 • Plot Statements

the statement must be binned by both X and Y. That is, the values for the X column and
the Y column must form a complete, rectangular grid of bins. Input data with non-binned
columns should be preprocessed with the KDE procedure (SAS/STAT), which enables
you to set the number of bins for X and Y. Alternatively, the data can be preprocessed
with a technique similar to the example, where a pre-defined bin width is used.
The bounding cube can be titled, rotated, and zoomed to provide a different viewpoint.
By default, the outline of the bounding cube is displayed and the viewing rotation angle
is 57 degrees, the tilt angle is 20 degrees and the zoom factor is 1. See the CUBE=,
ROTATE=, TILT=, and ZOOM= options of the LAYOUT OVERLAY3D statement for
information about how to change the viewpoint.
The X-axis, Y-axis, and Z-axis are linear by default. You can change axis properties with
the XAXISOPTS=, YAXISOPTS=, and ZAXISOPTS= options of the LAYOUT
OVERLAY3D statement.
Note: When BINAXIS=TRUE, some axis options for the X- and Y-axes might not
apply.

Example: BIHISTOGRAM3DPARM Statement

The following graph was generated by the “Example Program” on page 288:

Example Program
proc template;
define statgraph bihistogram;
begingraph;
entrytitle "Distribution of Height and Weight";
entryfootnote halign=right "SASHELP.HEART";
BLOCKPLOT Statement 289

layout overlay3d / cube=false zaxisopts=(griddisplay=on);

bihistogram3dparm x=height y=weight z=count / display=all;
endlayout;
endgraph;
end;
run;

data heart;
set sashelp.heart(keep=height weight);
if height ne . and weight ne .;
height=round(height,5);
weight=round(weight,25);
run;

proc summary data=heart nway completetypes;

class height weight;
var height;
output out=stats(keep=height weight count) N=Count;
run;

proc sgrender data=stats template=bihistogram;

run;

BLOCKPLOT Statement
Creates one or more strips of rectangular blocks containing text values. The width of each block
corresponds to specified numeric intervals.

Syntax
BLOCKPLOT X=column | expression
BLOCK=column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
ALTFILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of alternate fills.
BLOCKINDEX=positive-integer-column | expression
specifies indices for remapping the assignment of the color of the block fills.
CLASS=column | expression
creates a separate block plot for each unique value of the specified column or
expression.
DATATRANSPARENCY=number
specifies the degree of the transparency of the block fill and outline.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display an outlined, a filled, or an outlined and filled
block area.
EXTENDBLOCKONMISSING=TRUE | FALSE
specifies whether a missing value in the BLOCK column starts a new block
or reverts to the previous nonmissing value.
290 Chapter 6 • Plot Statements

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the block fills.
FILLTYPE=MULTICOLOR | ALTERNATE
specifies how the blocks are filled.
INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether missing values in the class column are included in the plot.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the block outlines.
REPEATEDVALUES=TRUE | FALSE
specifies whether contiguous block values that are identical create separate
blocks.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Label options
BLOCKLABEL=column | expression
specifies alternative text to display for the internal block text values.
LABEL="string"
specifies an external label for a single block plot.
LABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the external block label(s).
LABELPOSITION=LEFT | RIGHT | TOP | BOTTOM
specifies the alignment of BLOCK label.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Text options
VALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the appearance of the internal block text values or the alternative
values that are specified by the BLOCKLABEL= option.
VALUEFITPOLICY=NONE | SHRINK | SPLIT | SPLITALWAYS | TRUNCATE
specifies how text values are adjusted to fit within the containing block.
VALUEHALIGN=LEFT | CENTER | RIGHT | START
specifies the horizontal alignment of the value text within the blocks.
VALUESPLITCHAR="character-list"
specifies one or more characters on which the values can be split, if needed.
VALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed values.
VALUEVALIGN=TOP | CENTER | BOTTOM
specifies the vertical alignment of the value text within the blocks.

Required Arguments
X=column | expression
specifies numeric X axis positions. When the X-axis is numeric and the specified
column is numeric, values are expected to be in sorted, ascending order. If the X-axis
BLOCKPLOT Statement 291

is discrete and the specified column is numeric, values are treated as numeric-
discrete.
BLOCK=column | expression
specifies a value for each X position. Numeric values are converted to text strings
according to an assigned format or BEST6.

Optional Arguments
ALTFILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of alternate fills. This option in conjunction with the
FILLATTRS= option controls fill appearance when FILLTYPE=ALTERNATE.

Default The GraphWalls style element.

Requirement FILLTYPE= ALTERNATE must set for this option to have any
effect.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Tips The FILLATTRS= option controls the fill color.

To make all block fill areas the same color, set the FILLATTRS= and
ALTFILLATTRS= options to the same value.

The DATATRANSPARENCY= option sets the transparency for the

block fills and the outlines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines
but a different transparency for the alternate block fills. Example:
datatransparency=0.2 altfillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Fill Options” on page 1348 for available fill-options.

BLOCKINDEX=positive-integer-column | expression
specifies indices for remapping the assignment of the color of the block fills.

Requirements FILLTYPE= MULTICOLOR must be set for this option to have any
effect.

The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indices for a specific block value must be the same.
Otherwise, the results are unpredictable.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style element to use.
292 Chapter 6 • Plot Statements

If this option is not used, then the color values are mapped in the
order of GraphData1–GraphDataN.

BLOCKLABEL=column | expression
specifies alternative text to display for the internal block text values.

Default The BLOCK= values

Interaction The DISPLAY= option must include VALUES for the alternative text
to appear.

Note The text for each block segment must be the same. Otherwise, the
results are unpredictable.

Tips This option is particularly useful for showing regimes in forecasting

and time series plots.

The font and color attributes for the alternative text are specified by the
VALUEATTRS= option.

CLASS=column | expression
creates a separate block plot for each unique value of the specified column or
expression. Each block plot is labeled externally by the class value.

Interactions The DISPLAY= option must include LABEL for any external labels to
appear.

This option overrides the LABEL= option.

Tip The font and color attributes for the external labels are specified by the
LABELATTRS= option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the block fill and outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the block values or labels.

Tip The ALTFILLATTRS= option can be used to set transparency for just the
alternate block fills. The FILLATTRS= option can be used to specify
transparency for the block fills. You can combine this option with
ALTFILLATTRS= and with FILLATTRS= to set one transparency for the
outlines but a different transparency for the block fills. Example:
datatransparency=0.2
altfillattrs=(transparency=0.6) fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display an outlined, a filled, or an outlined and filled block area.
Values and a label can also be added or suppressed.
STANDARD
does the following based on the parent layout type:
• when used inside an overlay-type layout (OUTLINE FILL), STANDARD
displays an outlined, filled block without text values or a label.
BLOCKPLOT Statement 293

• when used as a stand-alone plot inside a LATTICE or GRIDDED layout

(OUTLINE FILL VALUES LABEL), STANDARD displays an outlined,
filled block with text values and a label.
ALL
displays all possible features.
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays an outlined, non-filled block

FILL displays a filled, non-outlined block
VALUES displays internal block values or the alternative block values
that are specified by the BLOCKLABEL= option
LABEL displays the external block label or labels

Default STANDARD

Tips Use the OUTLINEATTRS=, FILLATTRS=, ALTFILLATTRS=, and

BLOCKINDEX= options to control the appearance of the blocks.

Use the VALUEATTRS= and LABELATTRS= options to control the text

appearance.

EXTENDBLOCKONMISSING=TRUE | FALSE
specifies whether a missing value in the BLOCK column starts a new block or
reverts to the previous nonmissing value.

Default FALSE

Tip When EXTENDBLOCKONMISSING=TRUE, you can set up the input

data for the BLOCK= column with nonmissing values where you expect
the blocks to change and leave the remaining block values missing. For an
example, see “Example 1: BlockPlot Overlaid with SeriesPlot” on page
301.

See “boolean ” on page 1339 for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the block fills. This option in conjunction with the
ALTFILLATTRS= option controls fill appearance when FILLTYPE= ALTERNATE.

Default The GraphData1 style element.

Requirement FILLTYPE= ALTERNATE must set for this option to have any
effect.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Tips The ALTFILLATTRS= option controls the alternating fill color.

To make all block fill areas the same color, set the FILLATTRS= and
ALTFILLATTRS= options to the same value.
294 Chapter 6 • Plot Statements

The DATATRANSPARENCY= option sets the transparency for the

block fills and the outlines. You can combine this option with
TRANSPARENCY= in (fill-options) to set one transparency for the
outlines and a different transparency for the block fills. For example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Fill Options” on page 1348 for available fill-options.

FILLTYPE=MULTICOLOR | ALTERNATE
specifies how the blocks are filled.
MULTICOLOR
fills the blocks with the color specified by the COLOR attribute of the style
elements GraphData1–GraphDataN and GraphMissing (for missing BLOCK=
values), or the style elements indicated by the BLOCKINDEX= column.
ALTERNATE
alternates the block fill between the colors specified by the FILLATTRS= and
ALTFILLATTRS= options.

Default MULTICOLOR

Interactions For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

When this option is set to ALTERNATE, the block plot does not
support a DISCRETELEGEND entry.

INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether missing values in the class column are included in the plot.
Missing class values are included by default. When the data contains missing class
values, the label for those values is either blank for missing character values or a dot
for missing numeric values.
The following figure shows block plots for classes Class 1, Class 2, and any missing
class values.

Notice that the label for the missing class values is blank. You can use the
INCLUDEMISSINGCLASS=FALSE option to exclude the missing class values. If
you want to keep the missing class values, then you can create a format that specifies
a more meaningful label for the missing class. For example, here is a format that
specifies a label for missing character and numeric class values.
proc format;
value $missingClass " " = "(Missing)";
value missingClass . = "(Missing)";
run;
BLOCKPLOT Statement 295

A single space enclosed in quotation marks specifies a missing character value and a
dot specifies a missing numeric value. Although it might seem appropriate to use
empty quotation marks ('' or "") to specify a missing character value, doing so
produces unexpected results. To specify a missing character value, enclose a single
space in quotation marks (' ' or " "). You can use this format for the class columns in
the PROC SGRENDER statement. In that case, if the class columns contain missing
values, then the labels specified in the format statement are used for the missing
classes.
The following figure shows the previous example when format $missingClass is
applied to the class variable.

Note: In the second maintenance release of SAS 9.4 and in earlier releases, ODS
Graphics does not support Unicode values in user-defined formats. Starting with
the third maintenance release of SAS 9.4, ODS Graphics supports Unicode
values in user-defined formats only if they are preceded by the (*ESC*) escape
sequence. Example: "(*ESC*){unicode beta}". ODS Graphics does not
support an escape character that is defined in an ODS ESCAPECHAR statement
in user-defined formats.

Default TRUE

Interaction The CLASS= option must be specified for this option to have any
effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

LABEL="string"
specifies an external label for a single block plot.

Defaults The label of the BLOCK= column.

The name of the BLOCK= column, if there is no column label.

Interactions If the CLASS= option is specified, then this option is ignored.

The DISPLAY= option must include LABEL for any external label(s)
to appear.

Tip The font and color attributes for the external label are specified by the
LABELATTRS= option.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the external block label(s).

Default The GraphLabelText style element.

Interaction If one or more text options are specified and they do not include all of
the font properties such as color, family, size, weight, style, then the
non-specified properties are derived from the GraphLabelText style
element.
296 Chapter 6 • Plot Statements

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LABELPOSITION=LEFT | RIGHT | TOP | BOTTOM

specifies the alignment of BLOCK label.
LEFT
places the label to the left of the strip of block values.
RIGHT
places the label to the right of the strip of block values.
TOP
places the label above the strip of block values.
BOTTOM
places the label below the strip of block values.

Default LEFT

NAME="string"
assigns a name to this plot statement for reference in other template statements. This
option is used mostly in the DISCRETELEGEND statement in order to coordinate
the use of colors and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the block outlines.

Default The GraphOutlines style element.

Interactions For this option to have any effect, the outlines must be enabled by the
ODS style or the DISPLAY= option.

If labels are displayed in the TOP or BOTTOM position, then they are
also outlined.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

REPEATEDVALUES=TRUE | FALSE
specifies whether contiguous block values that are identical create separate blocks.
FALSE
creates only one block when two or more identical block values appear
consecutively.
TRUE
creates one block for each identical value when two or more identical values
appear consecutively.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.
BLOCKPLOT Statement 297

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the appearance of the internal block text values or the alternative values that
are specified by the BLOCKLABEL= option.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all of
the font properties such as color, family, size, weight, style, then the
non-specified properties are derived from the GraphValueText style
element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

VALUEFITPOLICY=NONE | SHRINK | SPLIT | SPLITALWAYS | TRUNCATE

specifies how text values are adjusted to fit within the containing block.
NONE
makes no attempt to fit values that collide.
SHRINK
reduces the font size of the values until they all fit.
SPLIT
splits a value that does not fit within the containing block at a split character in
order to make the value fit the available space. No split occurs at split characters
that occur where a split is not needed. If the value does not contain any of the
specified split characters, then a split does not occur.

Interaction The VALUESPLITCHAR= option specifies the split characters.

Tip Use the VALUEHALIGN= option to control the alignment of the

split lines.

SPLITALWAYS
splits each value at every occurrence of a split character. If the value does not
contain any of the specified split characters, then a split does not occur.

Interaction The VALUESPLITCHAR= option specifies the split characters.

Tip Use the VALUEHALIGN= option to control the alignment of the

split lines.

TRUNCATE
truncates any value that does not fit. For a numeric column, an asterisk (*) is
substituted for the entire value whenever truncation occurs. For a character
column, the truncated portion of the text is replaced by an ellipsis (...).

Default TRUNCATE

Interaction The SPLIT and SPLITALWAYS policies are ignored when the
BLOCKPLOT statement is placed in a DATALATTICE or
DATAPANEL layout and the BLOCKPLOT CLASS= option is set. In
that case, the TRUNCATE fit policy is used instead.

VALUEHALIGN=LEFT | CENTER | RIGHT | START

specifies the horizontal alignment of the value text within the blocks.
298 Chapter 6 • Plot Statements

LEFT
left-aligned within the block
CENTER
center-aligned within the block
RIGHT
right-aligned within the block
START
center-aligned at the starting value of the block

Default LEFT

Restriction When the BLOCKPLOT statement is placed inside an

INNERMARGIN block, only VALUEALIGN=CENTER is honored.

Requirement For this option to have any effect, the DISPLAY= option must
include VALUE.

Interaction When REPEATEDVALUES= TRUE and X values are numeric, only

CENTER and START can be used for a discrete axis. In addition,
only LEFT and START can be used for a linear or log axis. For
example, if REPEATEDVALUES=TRUE and the axis is discrete,
then a setting of RIGHT for this option is ignored and LEFT is used
instead.

VALUESPLITCHAR="character-list"
specifies one or more characters on which the values can be split, if needed. When
multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
value. In that case, all of the specified split characters together are treated as a single
split character.
When VALUESPLITPOLICY=SPLIT, if a value collision is detected, then the value
is split at each occurrence of any of the specified split characters. When
VALUESPLITPOLICY=SPLITALWAYS, the value is split unconditionally on each
of the specified split characters. If the value does not contain any of the specified
split characters, then the value is not split.
"character-list"
one or more characters with no delimiter between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the following
option:
valuesplitchar="abc"

The VALUEFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interactions The VALUESPLITCHARDROP= option specifies whether the split

characters are included in the displayed value or are dropped.
BLOCKPLOT Statement 299

The VALUEFITPOLICY= option sets the policy that is used to

manage the split behavior of the value.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

VALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed values.
TRUE
drops a split character from the value when a split occurs at that character. Split
characters at which a split does not occur are left in place. The
VALUEFITPOLICY= option determines where the values are split. When
VALUEFITPOLICY=SPLIT, each value is split at a split character only where a
split is needed in order to make the value fit the available space. At each split
point, the split character is dropped, and the characters that follow the split
character, up to but not including the split character at the next split point, are
wrapped to the following line.
When VALUEFITPOLICY=SPLITALWAYS, each value is split at every
instance of a split character. All of the split characters are dropped. The
characters that follow each split character, up to but not including the next split
character, are wrapped to the next line.
The following figure shows how value Product*Group*1 is split when the
VALUESPLITCHARDROP=TRUE, VALUESPLITCHAR="*", and
VALUEHALIGN=CENTER options are specified with the SPLIT and
SPLITALWAYS fit policies.

In this example, when VALUEFITPOLICY=SPLIT, the value is split at the first

occurrence of the asterisk in order to make the value fit. No split is needed at the
second asterisk. The first asterisk is dropped, and Group*1 wraps to the next line.
Notice that the second asterisk is not dropped in this case. When
VALUEFITPOLICY=SPLITALWAYS, the value is split at every occurrence of
the asterisk. In this case, both asterisks are dropped, and the characters that
follow each asterisk wrap to the next line.
FALSE
includes the split characters in the value display. The VALUEFITPOLICY=
option determines how the split characters are displayed. When
VALUEFITPOLICY=SPLIT, each value is split at a split character only where a
split is needed in order to make the value fit the available space. A split might not
occur at every split character in the value. At each split point, the split character
remains as the last character in the current line. The characters that follow the
split character, up to and including the split character at the next split point, are
then wrapped to the following line. This process repeats until the entire value is
displayed.
300 Chapter 6 • Plot Statements

When VALUEFITPOLICY=SPLITALWAYS, each value is split at every

instance of a split character in the value regardless of whether a split is actually
needed. Each split character remains as the last character in the current line. The
characters that follow each split character, up to and including the next split
character, are then wrapped to the next line.
The following figure shows how value Product*Group*1 is split when the
VALUESPLITCHARDROP=FALSE, VALUESPLITCHAR="*", and
VALUEHALIGN=CENTER options are specified with the SPLIT and
SPLITALWAYS fit policies.

In this example, when VALUEFITPOLICY=SPLIT, the label is split at the first

occurrence of the asterisk to make the value fit. No split is needed at the second
asterisk. The characters that follow the first asterisk wrap to the next line. When
VALUEFITPOLICY=SPLITALWAYS, the value is split at every occurrence of
the asterisk. Each asterisk remains as the last character in the current line, and the
characters that follow are wrapped to the next line.

Default TRUE. A split character is dropped from the value display when a
split occurs at that character.

Requirement The VALUEFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interaction The VALUESPLITCHAR= option specifies the split characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

VALUEVALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment of the value text within the blocks.

Default CENTER

Interaction For this option to have any effect, the DISPLAY= option must include
VALUE.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.
Example 1: BlockPlot Overlaid with SeriesPlot 301

Details
A block plot contains information about X locations and text values to be associated
with corresponding intervals along the X axis. There is no Y axis information in this
plot.

Examples

Example 1: BlockPlot Overlaid with SeriesPlot

When overlaid with plots that have a Y axis (a series plot for example), a block plot
expands vertically to fill the Y axis range. In the BLOCKPLOT statement, the BLOCK=
argument can be used to reference text values from a column. The resulting graph
displays those values within the plot wall. This example shows how a block plot can be
specified with a series plot within an OVERLAY layout. Here is the output for this
example.

Example Output

Example Program
data MSevents;
input Date date9. Release $5.;
label Release="Windows Release";
datalines;
01jun1990 3.0
01sep1995 95
01jul1998 98
01mar2000 2000
01nov2001 XP
;
302 Chapter 6 • Plot Statements

proc sort data=sashelp.stocks(keep=date stock close)

out=MSstock;
where stock="Microsoft";
by date;
run;

data events;
merge MSstock MSevents;
by date;
run;

proc template;
define statgraph blockplot1;
begingraph;
entrytitle "Microsoft Share Prices";
entrytitle "and Significant OS Releases";
layout overlay;
blockplot x=date block=release /
datatransparency=0.3 valuevalign=top
labelposition=top display=(fill values label)
extendblockonmissing=true ;
seriesplot x=date y=close;
endlayout;
endgraph;
end;
run;

proc sgrender data=events template=blockplot1;

format date year4.;
label date="Year";
run;

Program Description

Prepare the data. To prepare data for the graph, “event” information must be added to
the existing data for stock prices. Here is the SAS code that generates the data for this
example. Notice that the first DATA step creates a Release column. That column is later
specified on the BLOCK= argument to display text values on the wall of the block plot.
For more information about the merged input data, see “About the Merged Data” on
page 303.
data MSevents;
input Date date9. Release $5.;
label Release="Windows Release";
datalines;
01jun1990 3.0
01sep1995 95
01jul1998 98
01mar2000 2000
01nov2001 XP
;

proc sort data=sashelp.stocks(keep=date stock close)

out=MSstock;
where stock="Microsoft";
by date;
Example 1: BlockPlot Overlaid with SeriesPlot 303

run;

data events;
merge MSstock MSevents;
by date;
run;

Define the template. In the GTL template code, BLOCK=RELEASE is specified in the
BLOCKPLOT statement so that the RELEASE values are displayed on the wall of the
resulting block plot. In this template, the BLOCKPLOT statement sets
EXTENDBLOCKONMISSING= TRUE so that missing values in the data revert to the
previous nonmissing value in the block plot. Thus, in the block plot, values are missing
until 01JUN90, when the value changes from missing to 3.0. The block plot retains that
3.0 value for subsequent observations until the next nonmissing value replaces it (in
this case, the value95 on 01SEP95). In the example output shown in “Example Output”
on page 301, the fill color for each of the nonmissing values is determined by the style
elements GraphData1–GraphData5 in this case. The fill color for the missing values is
determined by the next available GraphDataN style element, which is GraphData6 in this
case.
proc template;
define statgraph blockplot1;
begingraph;
entrytitle "Microsoft Share Prices";
entrytitle "and Significant OS Releases";
layout overlay;
blockplot x=date block=release /
datatransparency=0.3 valuevalign=top
labelposition=top display=(fill values label)
extendblockonmissing=true ;
seriesplot x=date y=close;
endlayout;
endgraph;
end;
run;

Generate the graph. Format the date values on the X axis as four-digit year values, and
label the X axis “Year.”
proc sgrender data=events template=blockplot1;
format date year4.;
label date="Year";
run;

About the Merged Data

In the merged input data set, the Release column value is missing for each observation
until the first event defined in the MSEvents data, 3.0, occurs in June of 1990. The
Release column value is missing again for the subsequent observations until the next
event defined in the MSEvents data, 95, occurs in September of 1995. Here is a sample
of the merged data.
Obs Stock Date Close Release
...
45 Microsoft 02APR90 $58.00
46 Microsoft 01MAY90 $73.00
47 Microsoft 01JUN90 $76.00 3.0
304 Chapter 6 • Plot Statements

48 Microsoft 02JUL90 $66.50

49 Microsoft 01AUG90 $61.50
50 Microsoft 04SEP90 $63.00
...

Example 2: Stand-Alone BlockPlot in Lattice Layout

When used as a stand-alone plot in a lattice layout, the block plot’s height can be
controlled. Using this technique, it is possible to include two or more “event” strips in a
plot. This example demonstrates this technique by creating a series plot with an event
strip as shown in the following figure.

Example Output

Example Program
In this example, the lattice ROWWEIGHTS=(0.04 0.96) option apportions 4% of the
vertical space to the block plot. Here is the code for this example.
proc template;
define statgraph blockplot2;
begingraph;
entrytitle "Microsoft Share Prices";
entrytitle "and Significant OS Releases";
layout lattice / rowweights=(0.04 0.96);
blockplot x=date block=release / datatransparency=0.3
valuefitpolicy=shrink labelposition=left
display=(fill label outline values)
extendblockonmissing=true;
seriesplot x=date y=close;
endlayout;
endgraph;
end;
BOXPLOT Statement 305

run;

proc sgrender data=events template=blockplot2;

format date year4.;
label date="Year";
run;

BOXPLOT Statement
Creates box plots that are computed from input data.
Tip: Starting with the third maintenance release of SAS 9.4, you can use subpixel
rendering with this statement. It is enabled by default. To disable subpixel rendering,
specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 33. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.

Syntax
BOXPLOT Y=numeric-column | expression </option(s)>;
BOXPLOT X=column | expression
Y=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
BOXWIDTH=number
specifies the width of a box as a ratio of the maximum possible width.
CAPSHAPE=SERIF | LINE | BRACKET | NONE
specifies the shape at the ends of the whiskers.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing
or bin width.
CONNECT=MEAN | MEDIAN | Q1 | Q3 | MIN | MAX
specifies that a connect line joins a statistic from box to box.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the lines connecting multiple boxes.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled boxes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the box outlines, box fill,
whiskers, mean, median, caps, connect lines, outliers, and data labels, if
displayed.
DISPLAY=STANDARD | ALL | (display-options)
specifies which additional features of the box plot to display.
EXTREME=TRUE | FALSE
specifies whether the whiskers can extend beyond the fences.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the boxes.
306 Chapter 6 • Plot Statements

INDEX=positive-integer-column | expression
specifies indices for mapping box attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.
INTERVALBOXWIDTH=AUTO | dimension
specifies the box width when an interval category (X) column is specified.
MEANATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the marker representing the mean within the box.
MEDIANATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the line representing the median within the box.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and of the boxes.
OUTLIERATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the markers representing the outliers.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the box outline.
SPREAD=TRUE | FALSE
specifies whether outliers with the same value are spread out to avoid
overlap.
WHISKERATTRS=style-element | style-element (line-options) | (line-options)
specifies the line properties of the whiskers and caps.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

OUTLIERTIP=(role-list)
specifies the information to display when the cursor is positioned over an
outlier.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a box
or whisker in the box plot.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column
specifies the labels of the outliers. Either a numeric or a character column can
be used.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the outlier labels.
DATALABELSPLIT=TRUE | FALSE
BOXPLOT Statement 307

specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label blocks.
LABELFAR=TRUE | FALSE
specifies whether all outliers or only far outliers are labeled.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all boxes from the discrete X ticks.
GROUP=column | discrete-attr-var | expression
creates a box plot for each unique group value of the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how to display the boxes that represent group values for the
coordinate pairs.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
DISPLAYSTATS=NONE | STANDARD | ALL | (statistics-list)
specifies the statistics to be displayed for each box.
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation
that is read.
PERCENTILE=1 | 2 | 3 | 4 | 5
specifies one of five definitions used to calculate percentiles.
WEIGHT=numeric-column | expression
specifies a column that contains a statistics calculation a priori weight for
each observation of the input data object.
WHISKERPERCENTILE=number
specifies the whisker length, in percentile units.

Required Arguments
Specifying only Y= creates a single box plot. Specifying both X= and Y= creates a box
plot for each unique value of X.
Y=numeric-column | expression
specifies the column for the Y values. This argument is required.
308 Chapter 6 • Plot Statements

X=column | expression
specifies the column for the X values. This argument is required if you want to create
a box plot for each unique X value.

Note For interval X values, if a user-defined format is applied to the X column, the
format should map each X value to only one unique formatted value.
Otherwise, unexpected results might occur.

Optional Arguments
BOXWIDTH=number
specifies the width of a box as a ratio of the maximum possible width.

Defaults For nongrouped data, the default is 0.4.

For grouped data, the default is 0.6.

Range 0–1, where 0 is the narrowest and 1 is the widest

Interactions For grouped box plots with a discrete X (category) axis, the box width
is a percentage of the CLUSTERWIDTH .

Prior to the third maintenance release of SAS 9.4, this option is

ignored for an interval box plot, and the box width is controlled by the
INTERVALBOXWIDTH= option. Starting with the third maintenance
release of SAS 9.4, this option is honored for an interval box plot, but
it can be overridden by the INTERVALBOXWIDTH= option.

CAPSHAPE=SERIF | LINE | BRACKET | NONE

specifies the shape at the ends of the whiskers.
SERIF
specifies a short line perpendicular to the whisker.
LINE
specifies a line perpendicular to the whisker that extends the width of the box.
BRACKET
specifies a line perpendicular to the whiskers that extends the width of the box
and that has short extensions at each end. The extensions are drawn in the
direction of the box.
NONE
specifies that no shape appears at the ends of the whiskers.
The following figure shows each of the shapes.

Default The GraphBox:CapStyle style reference.

BOXPLOT Statement 309

Interactions The cap color and the thickness are specified by the
WHISKERATTRS= option. The cap pattern is always solid.

The DISPLAY= option must include CAPS in order for cap lines to be
shown.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or bin
width.

Default 0.7

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.

Requirement For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

Note When the X axis is an interval axis, the cluster width is a fraction of
the smallest data interval.

CONNECT=MEAN | MEDIAN | Q1 | Q3 | MIN | MAX

specifies that a connect line joins a statistic from box to box.

Default The GraphBox:Connect style reference.

Requirement The DISPLAY= option must contain the CONNECT display-options

value for the connect line to be displayed.

Interaction This option applies only when the X= argument is used to generate
multiple boxes.

Note Starting with the third maintenance release of SAS 9.4, the connect
lines are drawn in axis order. In prior releases, they are drawn in data
order.
310 Chapter 6 • Plot Statements

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the lines connecting multiple boxes.

Default The GraphConnectLine style element.

Interactions If there is only one box, then this option is ignored.

If the DISPLAY= option does not include CONNECT, or if the

GROUP= option is used, then this option is ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

DATALABEL=column
specifies the labels of the outliers. Either a numeric or a character column can be
used.

Default No data labels are displayed

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See LABELFAR= option

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the outlier labels.

Default The GraphDataText style element.

Interactions This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

If one or more label options are specified and they do not include all
the font properties (color, family, size, weight, style), then the non-
specified properties are derived from the GraphDataText style
element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters. When set to
TRUE, the data label is split unconditionally at each occurrence of any of the
specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

See “boolean ” on page 1339 for other Boolean values that you can use.
BOXPLOT Statement 311

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
data label. In that case, all of the specified split characters together are treated as a
single split character.
When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is
split unconditionally at each occurrence of any of the specified split characters. If the
data label does not contain any of the specified characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
TRUE
drops the split characters from the data label.
FALSE
includes the split characters in the data label. When DATALABELSPLIT=TRUE
and DATALABELSPLITCHARDROP=FALSE, each split character remains as
the last character in the current line. The characters that follow the split character,
up to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a data label with the following
specifications:
• the data label text for this label is Product*Group*A
• DATALABELSPLIT=TRUE
• DATALABELSPLITCHARDROP=TRUE | FALSE
• DATALABELSPLITCHAR="*"
312 Chapter 6 • Plot Statements

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from the

label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains as the
last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

Default Justification is determined by the system.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled boxes. The following figure shows boxes
with each of the skins applied.
BOXPLOT Statement 313

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the box outlines, box fill, whiskers, mean,
median, caps, connect lines, outliers, and data labels, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
interior fill area of the boxes. You can combine this option with
FILLATTRS= to set one transparency for the box outlines and the
whiskers, mean, median, caps, and connect lines, but a different
transparency for the box fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEOFFSET=number
specifies an amount to offset all boxes from the discrete X ticks.

Default 0 (no offset, all boxes are centered on the discrete ticks)

Range -0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right when ORIENT=VERTICAL, and
up when ORIENT=HORIZONTAL. (If the layout's axis options set
REVERSE=TRUE, then the offset direction is also reversed.)

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

See “About the DISCRETEOFFSET= Option” on page 327

314 Chapter 6 • Plot Statements

DISPLAY=STANDARD | ALL | (display-options)

specifies which additional features of the box plot to display.
STANDARD
displays this combination of features (CAPS FILL MEAN MEDIAN
OUTLIERS)
ALL
displays all features
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:

CAPS displays caps at the ends of the whiskers

CONNECT displays the line connecting multiple boxes
FILL displays filled boxes
MEAN displays the mean symbol within the box
MEDIAN displays the median line within the box
NOTCHES displays notched boxes
OUTLIERS displays markers for the outliers
The endpoints of the notches are at the following computed locations:
1.58 IQR
median ± n

In the equation, IQR (IQR=Q3-Q1) is the interquartile range and n is the sample size.
The medians (central lines) of the two boxes are significantly different at
approximately the 0.05 level if the corresponding notches do not overlap.

Default The GraphBox:DisplayOpts style reference. If this style element does

not exist, then the default is STANDARD.

Interaction If EXTREME= TRUE, then the OUTLIERS feature is ignored

BOXPLOT Statement 315

Notes Starting with the third maintenance release of SAS 9.4, connect lines
are drawn in axis order. They are drawn in data order in prior releases.

Starting with the third maintenance release of SAS 9.4, when

DISPLAY= includes MEAN, the BOXPLOT statement contributes its
mean markers to a discrete legend when TYPE=MARKER is in effect
in the DISCRETELEGEND statement.

Tips To control the appearance of these features, use the

CONNECTATTRS= , FILLATTRS= , MEANATTRS= ,
MEDIANATTRS= , OUTLIERATTRS= , and WHISKERATTRS=
options. The WHISKERATTRS= option controls affects both CAPS
and WHISKERS.

Regardless of which display options are being displayed, this option

does not affect the axis range.

DISPLAYSTATS=NONE | STANDARD | ALL | (statistics-list)

specifies the statistics to be displayed for each box.
NONE
does not display any statistics.
STANDARD
displays N, MEAN, and STD.
ALL
displays all available statistics (see the statistics-list)
(statistics-list)
a space-separated list of one or more of the following statistics, enclosed in
parentheses:

DATAMAX maximum data value that includes not only the maximum
whisker values but also the maximum outlier values. This
option is valid in the first maintenance release of SAS 9.4 and
later releases. The DATAMAX value is greater than or equal
to the MAX value and is always represented in the axis
range.
DATAMIN minimum data value that includes not only the minimum
whisker values but also the minimum outlier values. This
option is valid in the first maintenance release of SAS 9.4 and
later releases. The DATAMIN value is less than or equal to
the MIN value and is always represented in the axis range.
IQR interquartile range (Q3–Q1).
MAX maximum data value below the box upper fence.
MEAN mean data value for the box.
MEDIAN median data value for the box.
MIN minimum data value above the box lower fence.
N number of observations for the box.
Q1 lower quartile (25th percentile) for the box.
Q3 upper quartile (75th percentile) for the box.
RANGE range of the data (MAX–MIN).
316 Chapter 6 • Plot Statements

STD standard deviation of the data for the box.

SUMWGT sum of the weights for the box. This option is valid in the
first maintenance release of SAS 9.4 and later releases.

Default NONE

Restriction This option is ignored if ORIENT= HORIZONTAL

Note The notches in the box plot can extend beyond DATAMIN and
DATAMAX in some cases.

EXTREME=TRUE | FALSE
specifies whether the whiskers can extend beyond the fences. Fences are locations
above and below the box. The upper and lower fences are located at a distance 1.5
times the Interquartile Range (IQR) (IQR = Q3 - Q1). The upper and lower far
fences are located at a distance 3 times the IQR .
FALSE
specifies that whiskers be drawn from the upper edge of the box to the largest
value within the upper fence, and from the lower edge of the box to the smallest
value within the lower fence. This representation is sometime called a schematic
box plot or a Tukey box plot.
TRUE
specifies that whiskers be drawn to the largest and smallest data values, whether
these values are inside or outside the fences. The outliers and far outliers are not
displayed and are not labeled. This representation is sometime called a skeletal
box and whisker plot.

Default FALSE

Interaction This option overrides the DATALABEL= , DATALABELATTRS= ,

LABELFAR= , OUTLIERATTRS= , and SPREAD= options.

See “Statement Summary” on page 326

“boolean ” on page 1339 for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the boxes.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

For grouped data, the Color attribute of the GraphData1–GraphDataN

style elements.

Interaction For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

Tip The DATATRANSPARENCY option sets the transparency for the box
outlines, box fill, whiskers, mean, median, caps, connect lines, and
outliers. You can combine this option with DATATRANSPARENCY=
to set one transparency for the box outlines and the whiskers, mean,
median, caps, and connect lines, but a different transparency for the box
fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
BOXPLOT Statement 317

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that is
read.

Default All observations have a frequency count of 1.

Restriction If the value of the numeric-column is missing or is less than 1, then the
observation is not used in the analysis. If the value is not an integer,
then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation.

GROUP=column | discrete-attr-var | expression

creates a box plot for each unique group value of the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

For example, the Sashelp.Cars data that is used in the “Example Program” on page
336 contains a column named Origin, which identifies the region that produces each
car. This column could be used in the BOXPLOT statement to group the box plots in
the display (see the GROUPDISPLAY= option to see the output for the grouped
boxes):
layout overlay / xaxisopts=(display=(line ticks tickvalues));
boxplot y=mpg_city x=cylinders / name="b"
datalabel=make spread=true
display=(caps fill mean median)
group=origin ;
discretelegend "b" / title="Vehicle Type: ";
endlayout;

Defaults Each distinct group value is represented in the plot by a different box
outline color. The outline colors are defined by the
GraphData1:ContrastColor–GraphDataN:ContrastColor style
references.

If the MEDIAN, CAPS, or NOTCHES are enabled by the DISPLAY=

option, then each of these features uses the same color as the box
outline. Line styles do not change by group value.

If the MEAN or OUTLIERS are enabled by the DISPLAY= option,

then each distinct group value is represented by a different marker.
The markers are defined by the MarkerSymbol and ContrastColor
attributes of the GraphData1–GraphDataN and GraphMissing style
318 Chapter 6 • Plot Statements

elements. A marker is used for both MEAN and OUTLIERS, if

displayed.

If box fills are enabled by the ODS style or by the DISPLAY= option,
then each distinct group value is represented in the plot by a different
fill color. The fill colors are defined by the Color attribute of the
GraphData1–GraphDataN and GraphMissing style elements.

Interactions Connect lines are not drawn for grouped data.

The box plot display depends on the setting for the

GROUPDISPLAY= option.

By default, the group values are mapped in the order of the data. The
GROUPORDER= option can be used to control the sorting order of
the group values. The INDEX= option can be used to alter the default
sequence of colors and markers.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Tip The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the box outlines, but the
PATTERN= setting in the OUTLIERATTRS= option could be used to
assign the same line pattern to all box outlines and connect lines.

See “DISCRETEATTRVAR Statement” on page 1297

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how to display the boxes that represent group values for the coordinate
pairs. The following example shows a box plot with GROUPDISPLAY=CLUSTER:

OVERLAY
draws boxes for a given group value at the exact coordinate. Depending on the
data, boxes at a given coordinate might overlap.
CLUSTER
draws boxes for a given group value adjacent to each other.
BOXPLOT Statement 319

Default OVERLAY

Interaction This option is ignored unless GROUP= is specified.

Tip Use the CLUSTERWIDTH= option to control the width of the clusters
when CLUSTER is in effect.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
DESCENDING
orders the groups within a category in descending order.

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
320 Chapter 6 • Plot Statements

are determined by a GraphData1–GraphDataN style element instead of

by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping box attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

INTERVALBOXWIDTH=AUTO | dimension
specifies the box width when an interval category (X) column is specified.
AUTO
prior to the third maintenance release of SAS 9.4, AUTO uses 85% of the
smallest interval between any two boxes for the given plot. Starting with the third
maintenance release of SAS 9.4, AUTO uses the BOXWIDTH= option setting.
dimension
sets the box width to the specified value.

Default AUTO

Restriction The axis type for the category axis (X by default) must be LINEAR,
and the X column must be numeric.

Interaction Prior to the third maintenance release of SAS 9.4, this option controls
the box width for an interval box plot. Starting with the third
maintenance release of SAS 9.4, this option overrides the
BOXWIDTH= option for an interval box plot.

See “dimension” on page 1340

LABELFAR=TRUE | FALSE
specifies whether all outliers or only far outliers are labeled.
BOXPLOT Statement 321

FALSE
applies the labels specified by the DATALABEL= option to both the outliers and
the far outliers.
TRUE
applies the labels specified by the DATALABEL= option to the far outliers.

Default FALSE

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “Statement Summary” on page 326 for information about outliers.

“boolean ” on page 1339 for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y= column label. If a label is not defined, then the Y= column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

MEANATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the marker representing the mean within the box.

Defaults For non-grouped data, GraphBoxMean style element.

For grouped data, the MarkerSymbol, Markersize, and ContrastColor

attributes of the GraphData1–GraphDataN style elements.

Interaction This option is ignored if the DISPLAY= option does not display the
mean.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Marker Options” on page 1350 for available marker-options.

MEDIANATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the line representing the median within the box.

Defaults For non-grouped data, the GraphBoxMedian style element.

For grouped data, the LineStyle and LineThickness attritubes of the

GraphBoxMedian style element, and the ContrastColor attribute of the
GraphData1–GraphDataN style elements.

Interaction This option is ignored if the DISPLAY= option does not display the
median.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.
322 Chapter 6 • Plot Statements

“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and of the boxes.

Default VERTICAL

OUTLIERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the markers representing the outliers.

Defaults For non-grouped data, GraphOutlier style element.

For grouped data, the MarkerSymbol, Markersize, and ContrastColor

attributes of the GraphData1–GraphDataN style elements.

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Marker Options” on page 1350 for available marker-options.

OUTLIERTIP=(role-list)
specifies the information to display when the cursor is positioned over an outlier. If
this option is used, then it replaces all of the information that is displayed by default.
Roles for columns that do not contribute to the box plot can be specified along with
roles that do contribute.
(role-list)
an ordered, space-separated list of unique BOXPLOT roles. BOXPLOT roles for
OUTLIERTIP include X, Y, STAT, and DATALABEL.

Note In the data tip, the STAT role displays the text “outlier” or “far
outlier” as applicable.

Example The following example displays data tips only for the column that is
assigned to the X role:
OUTLIERTIP=(X)

Default The columns assigned to these roles are automatically included in the
data tip information: X and Y .

Requirement To generate data tips, you must include an ODS GRAPHICS ON

statement that has the IMAGEMAP option specified, and write the
graphs to the ODS HTML destination.
BOXPLOT Statement 323

Interaction The labels and formats for the OUTLIERTIP variables can be
controlled with the TIPLABEL= and TIPFORMAT= options.

See the TIP= option for specifying the information to display when the
cursor is positioned over a box or whisker in the box plot.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the box outline.

Defaults For non-grouped data, the ContrastColor, LineThickness, and LineStyle

attributes of the GraphOutlines style element.

For grouped data and filled boxes, the LineStyle and LineThickness
attributes of the GraphOutlines style element, and the ContrastColor
attribute of the GraphData1–GraphDataN style elements.

For grouped data and unfilled boxes, the LineThickness attribute of the
GraphOutlines style element, and the ContrastColor and LineStyle
attributes of the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

PERCENTILE=1 | 2 | 3 | 4 | 5
specifies one of five definitions used to calculate percentiles.

Default 5 (empirical distribution function with averaging)

Note The percentile definition and default are the same as those that are used by
the PCTLDEF= option of the UNIVARIATE procedure and the
QNTLDEF= option of the SUMMARY procedure.

See “Calculating Percentiles” on page 329

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.
324 Chapter 6 • Plot Statements

SPREAD=TRUE | FALSE
specifies whether outliers with the same value are spread out to avoid overlap. For
vertical box plots this means offsetting the outliers horizontally. If this option is
false, then outliers with the same value are plotted in the same position, which means
only one is visible.

Default FALSE

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “boolean ” on page 1339 for other Boolean values that you can use.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a box or
whisker in the box plot. If this option is used, then it replaces all of the information
that is displayed by default.
(role-list)
an ordered, space-separated list of unique BOXPLOT roles. BOXPLOT roles for
TIP include DATAMAX, DATAMIN, MAX, MIN, MEAN, MEDIAN, N, Q1,
Q2, STD, SUMWGT, and X.

Note The roles DATAMAX, DATAMIN, and SUMWGT apply to the first
maintenance release of SAS 9.4 and later releases.

Tip Statistics such as N, MIN, and MAX are special roles, They are not
column-based like the X role.

Example The following example displays data tips only for the columns that
are assigned to the roles X (CATEGORY) and the statistic MEAN:
TIP=(X MEAN)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X, N, STD, MIN, MAX, Q1, Q3, MEAN, and
MEDIAN.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

See the OUTLIERTIP= option for specifying the information to display

when the cursor is positioned over an outlier.
BOXPLOT Statement 325

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIP=(X Y)
TIPFORMAT=(X=4. Y=4.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the OUTLIERTIP= or TIP= options are
used.

Requirement A column must be assigned to each of the specified roles.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example TIP=(X)
TIPLABEL=(X="Type")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the OUTLIERTIP= or TIP= options are
used.

Requirement A column must be assigned to each of the specified roles.

WEIGHT=numeric-column | expression
specifies a column that contains a statistics calculation a priori weight for each
observation of the input data object.

Requirement The value must be nonnegative.

Interaction If the value for an observation is missing or is less than 1, then the
observation is removed from the analysis.

WHISKERATTRS=style-element | style-element (line-options) | (line-options)

specifies the line properties of the whiskers and caps.

Defaults For non-grouped data, the GraphBoxWhisker style element.

For grouped data, the LineStyle and LineThickness attritubes of the

GraphBoxWhisker style element, and the ContrastColor attribute of the
GraphData1–GraphDataN style elements.

Restriction The caps are always drawn with a solid line.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.
326 Chapter 6 • Plot Statements

“Line Options” on page 1349 for available line-options.

WHISKERPERCENTILE=number
specifies the whisker length, in percentile units. When this option is specified,
number is used as the low percentile, and 100–number is used as the high percentile.
Here are some examples of values and their effect:

0 specifies the high and low extremes

10 specifies the 10th percentile low and the 90th percentile high
25 specifies the 25th percentile low and the 75th percentile high

Default The whiskers are drawn from the box to the most extreme point that is less
than or equal to 1.5 times the IQR

Range 0–25

Notes When this option is specified, fences and far outliers are not drawn.

When this option is set to 25, no whiskers are drawn because the box
extends from the 25th to the 75th percentile.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interactions This option is ignored if the X= argument is not specified.

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details

Statement Summary
The BOXPLOT statement displays a single box if given just a Y argument. It displays
multiple boxes if given both Y and X arguments and X has more than one unique value.
By default for numeric or character columns, the category (X) axis is
TYPE=DISCRETE. You can override the default and set the TYPE= to LINEAR or
TIME in the parent layout, provided that the X column is numeric. The axis for the
analysis (Y) column is always LINEAR. When the X axis is LINEAR, you must use the
INTERVALBOXWIDTH= option to specify the box width.
BOXPLOT Statement 327

When ORIENT= VERTICAL, the X (or X2) axis is used for the X column and the Y (or
Y2) axis is used for the Y column. When ORIENT=HORIZONTAL, the X (or X2) axis
is used for the Y column and the Y (or Y2) axis is used for the X column.
Two basic box plot representations can be drawn with the BOXPLOT statement: a
schematic (Tukey) box plot and a skeletal box plot. See the EXTREME= option for
details.
The following figure illustrates the box plot elements:

As shown in the figure, the bottom and top edges of the box are located at the 25th and
75th percentiles of the sample. Within the box, you can display the median (50th
percentile) as a line and the mean as a marker (see DISPLAY= option).
You can also display markers and data labels for outliers. Outliers are observations that
are more extreme than the upper and lower fences (± 1.5 IQR). Outliers that are beyond
upper and lower far fences (± 3 IQR) are called FAR OUTLIERS and can also be
identified and labeled. From a graphical perspective, the location of fences along the axis
are known, but there is no line or marker that displays a fence. (See DISPLAY= ,
LABELFAR= , and DATALABEL= options).
Finally, you can control the range represented by the whiskers. By default, the whiskers
are drawn from the upper edge of the box to the MAX value, and from the lower edge of
the box to the MIN value. (See the EXTREME= option.)

About the DISCRETEOFFSET= Option

This feature is useful for graphing multiple response variables side by side on a common
axis. By default within an overlay-type layout, if multiple BOXPLOT statements are
used with different analysis variables, then the boxes for matching X values are centered
on the ticks. Depending on the data, the boxes might be superimposed. The following
code fragment shows the default box positioning:
layout overlay / cycleattrs=true
328 Chapter 6 • Plot Statements

yaxisopts=(label="Miles Per Gallon");

boxplot x=type y=mpg_city / name="City"
legendlabel="City";
boxplot x=type y=mpg_highway / name="Highway"
legendlabel="Highway";
discretelegend "City" "Highway";
endlayout;

To place the different response values side by side, you can assign a different offset to
each BOXPLOT statement. The BOXWIDTH= option can be used in conjunction with
the DISCRETEOFFSET= option to create narrower boxes when desired.
layout overlay / cycleattrs=true
yaxisopts=(label="Miles Per Gallon");
boxplot x=type y=mpg_city / name="City"
discreteoffset=0.2 legendlabel="City";
boxplot x=type y=mpg_highway / name="Highway"
discreteoffset=-0.2 legendlabel="Highway";
discretelegend "City" "Highway";
endlayout;
BOXPLOT Statement 329

Calculating Percentiles
You can specify one of five definitions for computing the percentiles with the
PERCENTILE= option. Let n be the number of nonmissing values for a variable, and let
X 1, X 2, ..., X n represent the ordered values of the variable. X 1 is the smallest value, X 2 is
the next smallest, and X n is the largest value. Let the tth percentile be y, set:
t
p = 100

and let:
np = j + g

when PERCENTILE=1, 2, 3, or 5, or let:

(n + 1)p = j + g

when PERCENTILE=4, where j is the integer part of np, and g is the fractional part of
np. Then the PERCENTILE= option defines the tth percentile, y, as described in the
following table:

Percentile Description Equation Notes

1 Weighted average at Xnp y = (1 − g)x j + gx j + 1 x0 is taken to be x1

2 Observation numbered y = xj 1
Used when g < 2 or when
closest to np
1
g = 2 and j is even

y = xj + 1 1
Used when g = 2 and j is
1
odd or when g > 2
330 Chapter 6 • Plot Statements

Percentile Description Equation Notes

3 Empirical distribution y = xj Used when g = 0

function

y = xj + 1 Used when g > 0

4 Weighted average aimed y = (1 − g)x j + gx j + 1 xn + 1 is taken to be xn

at x(n + 1) p

5 Empirical distribution 1 Used when g = 0

y = 2 (x j + x j + 1)
function with averaging

y = xj + 1 Used when g > 0

Changing Box Plot Display

The SAS System defines graphical style elements that control the display of box plots
generated with the BOXLOT or BOXPLOTPARM statement. Using these style elements
as a starting point, you can change the style attribute values to achieve a very different
appearance for your box plots. Using the DEFAULT style for an example, here is a
portion of the style template for elements that are related to box plots:
proc template;
define style Default;
...
style GraphBox /
capstyle = "serif"
connect = "mean"
displayopts = "fill caps median mean outliers";
style GraphBoxMean / ...;
style GraphBoxMedian / ...;
style GraphBoxOutlier / ...;
style GraphBoxWhisker / ...;
...
end;
run;

Style Element Purpose

GraphBox general box plot properties (see the next table)

GraphBoxMean marker properties of mean marker

(MARKERSYMBOL= , MARKERSIZE=, CONTRASTCOLOR=)

GraphBoxMedian line properties of the median line

(LINESTYLE=, LINETHICKNESS=, CONTRASTCOLOR=)

GraphBoxOutlier marker properties of outliers

(MARKERSYMBOL=, MARKERSIZE=, CONTRASTCOLOR=)
BOXPLOT Statement 331

Style Element Purpose

GraphBoxWhisker line properties of whiskers and caps

(LINESTYLE=, LINETHICKNESS=, CONTRASTCOLOR=)

Attributes and Values for the GraphBox Style Element

Attribute Value(s) Description

CONNECT= "MEAN" | "MEDIAN" | "Q1" | "Q3" | statistic to connect with line

"MIN" | "MAX"

CAPSTYLE= "SERIF" | "LINE" | "BRACKET" shape at ends of whiskers

DISPLAYOPTS= "<CAPS> show caps at end of whiskers

<FILL> show filled boxes

<MEAN> show a marker for the mean

<MEDIAN> show a line for the median

<OUTLIERS> show markers for the outliers

<CONNECT> show line connecting same

statistic on multiple boxes

<NOTCHES>" show notched boxes

The DISPLAYOPTS attribute of GraphBox lists the general features to be displayed.

The following diagram shows the standard display for box plots, as defined by the
DEFAULT style. The keywords that are related to the appearance features are annotated:
332 Chapter 6 • Plot Statements

The two display options that are not the default are CONNECT (show connect lines) and
NOTCHES.

The STATISTICAL style is derived from the DEFAULT style and inherits the GraphBox
element from the parent DEFAULT style. The following code generates a box plot for
the STATISTICAL style:
/* Specify a path for the ODS output */
filename odsout "output-path";

proc template;
define statgraph boxplotdef;
begingraph;
entrytitle "Statistical Style";
layout overlay / xaxisopts=(label="Age" type=linear);
boxplot x=ageatstart y=cholesterol / intervalboxwidth=40;
endlayout;
endgraph;
end;

ods graphics / outputfmt=static;

ods _all_ close;
ods html path=odsout file="boxplot.html" style=statistical;

proc sgrender data=sashelp.heart template=boxplotdef;

where ageatstart between 50 and 55;
run;

ods html close;

ods html; /* Not required in SAS Studio */
BOXPLOT Statement 333

Here is the output.

For this example, we want to change the following attributes on the default box plot:
• By default, serif caps are displayed at the end of the fences. We want to remove those
caps from the fence lines.
• By default, the boxes are filled. We want to display empty, notched boxes.
• By default, the mean values are represented by hollow diamonds. We want to display
filled diamonds and slightly reduce their size.
• By default, the marker symbols for the outliers are hollow black circles. We want to
change the size and shape of the marker symbols, and again reduce their size.
To make these changes, we can derive a new style from the STATISTICAL style and set
the attributes that we want to change. Any attribute settings that we do not change are
inherited from the parent STATISTICAL style. The following style template effects the
desired changes:
proc template;
define style Boxplot;
parent = styles.statistical;
style GraphBox from GraphBox /
capstyle = "line"
displayopts = "caps median mean outliers notches";
style GraphBoxMean from GraphBoxMean /
markersymbol="diamondfilled"
contrastcolor=GraphColors("gcdata1")
markersize = 5px;
style GraphOutlier from GraphOutlier /
markersize = 5px
markersymbol = "x"
contrastcolor = GraphColors("gcdata2");
end;
run;
334 Chapter 6 • Plot Statements

Note the following:

• The DEFINE STYLE statement assigns the name BOXPLOT to our new style, and
sets the STATISTICAL style as the parent style.
• On the GraphBox style element, the CAPSTYLE= attribute is set to LINE, which
removes the serif caps from the end of the fences. The DISPLAYOPTS= attribute
drops the FILL value from the display list and adds the NOTCHES value; these
changes determine that the graph displays empty, notched boxes.
• On the GraphBoxMean style element, the marker symbol is changed to a filled
diamond and the marker size is reduced to 5 pixels (the default is 9 pixels). The
CONTRASTCOLOR= attribute is set to GCDATA1 (the default is GCDATA).
• On the GraphBoxOutlier style element, the marker symbol is changed to an X and
the marker size is reduced to 5 pixels (the default is 7 pixels). The
CONTRASTCOLOR= attribute is set to GCDATA2 (the default is GCOUTLIER).
The following code generates a box plot for the BOXPLOT style:
/* Specify a path for the ODS output */
filename odsout "output-path";

proc template;
define statgraph boxplotdef;
begingraph;
entrytitle "Boxplot Style";
layout overlay / xaxisopts=(label="Age" type=linear);
boxplot x=ageatstart y=cholesterol / intervalboxwidth=40;
endlayout;
endgraph;
end;

ods graphics / outputfmt=static;

ods _all_ close;
ods html path=odsout file="notchedboxplot.html" style=Boxplot;

proc sgrender data=sashelp.heart template=boxplotdef;

where ageatstart between 50 and 55;
run;

ods html close;

ods html; /* Not required in SAS Studio */
Example: BOXPLOT Statement 335

Here is the output.

When making such style changes remember that you are affecting all box plot displays
for all procedures that produce box plots when this style is in effect. It is possible to
change the box plot appearance for specific procedures, but to do this, a specific graph
template must be modified, not a style template.
For a comprehensive description of the style elements affecting ODS graphics, see
“Graph Style Elements Used by ODS Graphics” in SAS Graph Template Language:
User's Guide.

Example: BOXPLOT Statement

The following graph was generated by the “Example Program” on page 336:
336 Chapter 6 • Plot Statements

Example Program
proc template;
define statgraph boxplot;
begingraph;
entrytitle "City Mileage for Vehicle Types";
layout overlay
xaxisopts=(offsetmin=0.1 offsetmax=0.1);
boxplot y=mpg_city x=type /
datalabel=make spread=true;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=boxplot;

label type="Vehicle Type";
run;

BOXPLOTPARM Statement
Creates side-by-side box plots specified by parameters.
Requirements: The input data must be precomputed. See “Input Data Requirements for the
BOXPLOTPARM Statement” on page 362.
Nonmissing Y values for statistical observations of Q1 and Q3 are required for a box
to be drawn.
The statistical values, if present, must conform to the following rules for a box to be
displayed:
Q1 <= MEDIAN <= Q3
BOXPLOTPARM Statement 337

MIN <= MAX

STD >= 0
N > 0

Tip: Starting with the third maintenance release of SAS 9.4, you can use subpixel
rendering with this statement. It is enabled by default. To disable subpixel rendering,
specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 33. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.

Syntax
BOXPLOTPARM Y=numeric-column | expression
STAT=string-column </option(s)>;
BOXPLOTPARM X=column | expression
Y=numeric-column | expression
STAT=string-column </option(s)>;

Summary of Optional Arguments

Appearance options
BOXWIDTH=number
specifies the width of a box as a ratio of the maximum possible width.
CAPSHAPE=SERIF | LINE | BRACKET | NONE
specifies the shape at the ends of the whiskers.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing
or bin width.
CONNECT=MEAN | MEDIAN | Q1 | Q3 | MIN | MAX
specifies that a connect line joins a statistic from box to box.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the lines connecting multiple boxes.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled boxes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the box outlines, box fill,
whiskers, mean, median, caps, connect lines, outliers, and data labels, if
displayed.
DISPLAY=STANDARD | ALL | (display-options)
specifies which additional features of the box plot to display.
EXTREME=TRUE | FALSE
specifies whether the whiskers can extend beyond the fences.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the boxes.
INDEX=positive-integer-column | expression
specifies indices for mapping box attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.
INTERVALBOXWIDTH=AUTO | dimension
specifies the box width when an interval category (X) column is specified.
338 Chapter 6 • Plot Statements

MEANATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the marker representing the mean within the box.
MEDIANATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the line representing the median within the box.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and of the boxes.
OUTLIERATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the markers representing the outliers.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the box outline.
SPREAD=TRUE | FALSE
specifies whether outliers with the same value are spread out to avoid
overlap.
WHISKERATTRS=style-element | style-element (line-options) | (line-options)
specifies the line properties of the whiskers and caps.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

OUTLIERTIP=(role-list)
specifies the information to display when the cursor is positioned over an
outlier.
ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a box
or whisker in the box plot.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column
specifies the labels of the values that are identified as outlier or far outlier by
the STAT= column.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the outlier labels.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
BOXPLOTPARM Statement 339

specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label blocks.
LABELFAR=TRUE | FALSE
specifies whether all outliers or only far outliers are labeled.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all boxes from the discrete X ticks.
GROUP=column | discrete-attr-var | expression
creates a box plot for each unique group value of the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how to display the boxes that represent group values for the
coordinate pairs.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page that is displayed when a box or an outlier marker is
selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
DISPLAYSTATS=NONE | STANDARD | ALL | (statistics-list)
specifies the statistics to be displayed for each box.

Required Arguments
Specifying only Y= creates a single box plot. Specifying both X= and Y= creates a box
plot for each unique value of X.
Y=numeric-column | expression
specifies the column for the Y values. The Y values must be the statistical values
needed for the box plot. At a minimum, there must be nonmissing values for the 25th
and 75th percentiles.
X=column | expression
specifies the column for the X values. The X values must qualify or classify the
values in the Y column. This optional argument is used to create a plot box for each
classifier.
340 Chapter 6 • Plot Statements

Restriction When you specify a numeric column for the X= argument and the
category axis is an interval axis, you should not change the column
value format. Doing so might result in incorrect output.

STAT=string-column
specifies the statistic that is represented by the value in the Y column. Valid STAT=
column values include the following (see the requirements listed at the end of this
description):
BOXWIDTH
specifies the width of the boxes as a ratio of the maximum possible width. The
range of values is 0 (narrowest) to 1 (widest). The default is 0.4. If the Y value
corresponding to BOXWIDTH is in range, then it overrides the setting that is
specified in the BOXWIDTH= option.
DATAMAX
specifies the maximum data value that includes not only the maximum whisker
values but the maximum outlier values as well. The DATAMAX value is greater
than or equal to the MAX value and is always represented in the axis range.

Note This statistic is valid in the first maintenance release of SAS 9.4 and later
releases.

DATAMIN
specifies the minimum data value that includes not only the minimum whisker
values but the minimum outlier values as well. The DATAMIN value is less than
or equal to the MIN value and is always represented in the axis range.

Note This statistic is valid in the first maintenance release of SAS 9.4 and later
releases.

FAROUTLIER
specifies the observations that are outside the lower and upper far fences. The far
fences are located at a distance 3 times the Interquartile Range (IQR = Q3–Q1)
above and below the box. The far outliers are labeled when the DATALABEL=
option is used. Specify that LABELFAR= TRUE to label only the far outliers but
not the outliers.
MAX
specifies the maximum data value less than or equal to the upper fence.
MEAN
specifies the data mean.
MEDIAN
specifies the data median.
MIN
specifies the minimum data value greater than or equal to the lower fence.
N
specifies the subgroup sample size. The N value is not shown in the plot but is
used to calculate notch locations when the DISPLAY= option displays notches.
OUTLIER
specifies the observations that are outside the lower and upper fences. The fences
are located at a distance 1.5 times the Interquartile Range (IQR = Q3–Q1) above
and below the box. The outliers are labeled when the DATALABEL= option is
used.
BOXPLOTPARM Statement 341

Q1
specifies the 1st quartile (25th percentile). The data must contain a nonmissing
value for this quartile.
Q3
specifies the 3rd quartile (75th percentile). The data must contain a nonmissing
value for this quartile.
STD
specifies the data standard deviation.
SUMWGT
specifies the sum of the weights for the box.

Note This statistic is valid in the first maintenance release of SAS 9.4 and later
releases.

Requirements Nonmissing Y values for STAT observations of Q1 and Q3 are

required for a box to be drawn. Other STAT values can be omitted or
have missing Y values.

The STAT values, if present, must conform to the following rules for
a box to be displayed:
Q1 <= MEDIAN <= Q3
MIN <= MAX
STD >= 0
N > 0

Optional Arguments
BOXWIDTH=number
specifies the width of a box as a ratio of the maximum possible width.

Defaults For nongrouped data, the default is 0.4.

For grouped data, the default is 0.6.

Range 0–1, where 0 is the narrowest and 1 is the widest

Interactions For grouped box plots with a discrete X (category) axis, the box width
is a percentage of the CLUSTERWIDTH .

This option is overridden by the Y value when the STAT= column

value is BOXWIDTH and the corresponding Y value is in range.

Prior to the third maintenance release of SAS 9.4, this option is

ignored for an interval box plot, and the box width is controlled by the
INTERVALBOXWIDTH= option. Starting with the third maintenance
release of SAS 9.4, this option is honored for an interval box plot, but
it can be overridden by the INTERVALBOXWIDTH= option.

CAPSHAPE=SERIF | LINE | BRACKET | NONE

specifies the shape at the ends of the whiskers.
SERIF
specifies a short line perpendicular to the whisker.
342 Chapter 6 • Plot Statements

LINE
specifies a line perpendicular to the whisker that extends the width of the box.
BRACKET
specifies a line perpendicular to the whiskers that extends the width of the box
and that has short extensions at each end. The extensions are drawn in the
direction of the box.
NONE
specifies that no shape appears at the ends of the whiskers.
The following figure shows each of the shapes.

Default The GraphBox:CapStyle style reference.

Interactions The cap color and the thickness are specified by the
WHISKERATTRS= option. The cap pattern is always solid.

The DISPLAY= option must include CAPS in order for cap lines to be
shown.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or bin
width.

Default 0.7

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.
BOXPLOTPARM Statement 343

Requirement For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

CONNECT=MEAN | MEDIAN | Q1 | Q3 | MIN | MAX

specifies that a connect line joins a statistic from box to box.

Default The GraphBox:Connect style reference.

Requirement The DISPLAY= option must contain the CONNECT display-options

value for the connect line to be displayed.

Interaction This option applies only when the X= argument is used to generate
multiple boxes.

Note Starting with the third maintenance release of SAS 9.4, the connect
lines are drawn in axis order. In prior releases, they are drawn in data
order.

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the lines connecting multiple boxes.

Default The GraphConnectLine style element.

Interactions If there is only one box, then this option is ignored.

If the DISPLAY= option does not include CONNECT, or if the

GROUP= option is used, then this option is ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

DATALABEL=column
specifies the labels of the values that are identified as outlier or far outlier by the
STAT= column. Either a numeric or a character column can be used.

Default No data labels are displayed

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See the LABELFAR= option

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the outlier labels.

Default The GraphDataText style element.

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

344 Chapter 6 • Plot Statements

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters. When set to
TRUE, the data label is split unconditionally at each occurrence of any of the
specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
data label. In that case, all of the specified split characters together are treated as a
single split character.
When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is
split unconditionally at each occurrence of any of the specified split characters. If the
data label does not contain any of the specified characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
TRUE
drops the split characters from the data label.
BOXPLOTPARM Statement 345

FALSE
includes the split characters in the data label. When DATALABELSPLIT=TRUE
and DATALABELSPLITCHARDROP=FALSE, each split character remains as
the last character in the current line. The characters that follow the split character,
up to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a data label with the following
specifications:
• the data label text for this label is Product*Group*A
• DATALABELSPLIT=TRUE
• DATALABELSPLITCHARDROP=TRUE | FALSE
• DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from the

label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains as the
last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

Default Justification is determined by the system.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled boxes. The following figure shows boxes
with each of the skins applied.
346 Chapter 6 • Plot Statements

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the box outlines, box fill, whiskers, mean,
median, caps, connect lines, outliers, and data labels, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
interior fill area of the boxes. You can combine this option with
FILLATTRS= to set one transparency for the box outlines and the
whiskers, mean, median, caps, and connect lines, but a different
transparency for the box fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
BOXPLOTPARM Statement 347

DISCRETEOFFSET=number
specifies an amount to offset all boxes from the discrete X ticks.

Default 0 (no offset, all boxes are centered on the discrete ticks)

Range -0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right when ORIENT=VERTICAL, and
up when ORIENT=HORIZONTAL. (If the layout's axis options set
REVERSE=TRUE, then the offset direction is also reversed.)

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

See “About the DISCRETEOFFSET= Option” on page 364

DISPLAY=STANDARD | ALL | (display-options)

specifies which additional features of the box plot to display.
STANDARD
displays this combination of features (CAPS FILL MEAN MEDIAN
OUTLIERS)
ALL
displays all features
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:

CAPS displays caps at the ends of the whiskers

CONNECT displays the line connecting multiple boxes
FILL displays filled boxes
MEAN displays the mean symbol within the box
MEDIAN displays the median line within the box
NOTCHES displays notched boxes
OUTLIERS displays markers for the outliers
The endpoints of the notches are at the following computed locations:
1.58 IQR
median ±
n

In the equation, IQR is the interquartile range and n is the sample size.
348 Chapter 6 • Plot Statements

Default The GraphBox:DisplayOpts style reference. If this style element does

not exist, then the default is STANDARD.

Restriction The display features requested can be displayed only if the input data
includes this information.

Interaction If EXTREME= TRUE, then the OUTLIERS feature is ignored

Notes Starting with the third maintenance release of SAS 9.4, the connect
lines are drawn in axis order. They are drawn in data order in prior
releases.

Starting with the third maintenance release of SAS 9.4, when

DISPLAY= includes MEAN, the BOXPLOTPARM statement
contributes its mean markers to a discrete legend when
TYPE=MARKER is in effect in the DISCRETELEGEND statement.

Tips To control the appearance of these features, use the

CONNECTATTRS= , FILLATTRS= , MEANATTRS= ,
MEDIANATTRS= , OUTLIERATTRS= , and WHISKERATTRS=
options. The WHISKERATTRS= option controls affects both CAPS
and WHISKERS.

Regardless of which display options are being displayed, this option

does not affect the axis range.

DISPLAYSTATS=NONE | STANDARD | ALL | (statistics-list)

specifies the statistics to be displayed for each box.
NONE
does not display any statistics.
STANDARD
displays N, MEAN, and STD.
ALL
displays all available statistics (see the statistics-list)
BOXPLOTPARM Statement 349

(statistics-list)
a space-separated list of one or more of the following statistics, enclosed in
parentheses:

DATAMAX maximum data value that includes not only the maximum
whisker values but also the maximum outlier values. This
option is valid in the first maintenance release of SAS 9.4 and
later releases. The DATAMAX value is greater than or equal
to the MAX value and is always represented in the axis
range.
DATAMIN minimum data value that includes not only the minimum
whisker values but also the minimum outlier values. This
option is valid in the first maintenance release of SAS 9.4 and
later releases. The DATAMIN value is less than or equal to
the MIN value and is always represented in the axis range.
IQR interquartile range (Q3–Q1).
MAX maximum data value below the box upper fence.
MEAN mean data value for the box.
MEDIAN median data value for the box.
MIN minimum data value above the box lower fence.
N number of observations for the box.
Q1 lower quartile (25th percentile) for the box.
Q3 upper quartile (75th percentile) for the box.
RANGE range of the data (MAX–MIN).
STD standard deviation of the data for the box.
SUMWGT sum of the weights for the box. This option is valid in the
first maintenance release of SAS 9.4 and later releases.

Default NONE

Restrictions This option is ignored if ORIENT= HORIZONTAL

Only those statistics that are included in the STAT= column can be
displayed. RANGE requires both MAX and MIN to be included. IQR
requires both Q1 and Q3 to be included.

Note The notches in the box plot can extend beyond DATAMIN and
DATAMAX in some cases.

EXTREME=TRUE | FALSE
specifies whether the whiskers can extend beyond the fences. Fences are locations
above and below the box. The upper and lower fences are located at a distance 1.5
times the Interquartile Range (IQR) (IQR = Q3 - Q1). The upper and lower far
fences are located at a distance 3 times the IQR .
FALSE
specifies that whiskers be drawn from the upper edge of the box to the largest
value within the upper fence, and from the lower edge of the box to the smallest
value within the lower fence. This representation is sometime called a schematic
box plot or a Tukey box plot.
350 Chapter 6 • Plot Statements

TRUE
specifies that whiskers be drawn to the largest and smallest data values, whether
these values are inside or outside the fences. The outliers and far outliers are not
displayed and are not labeled. This representation is sometime called a skeletal
box and whisker plot.

Default FALSE

Interaction This option overrides the DATALABEL= , DATALABELATTRS= ,

LABELFAR= , OUTLIERATTRS= , and SPREAD= options.

See “Statement Summary” on page 361

“boolean ” on page 1339 for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the boxes.

Defaults For non-grouped data, the Color attribute of GraphDataDefault style

element.

For grouped data, the Color attribute of GraphData1–GraphDataN style

elements.

Interaction For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

Tip The DATATRANSPARENCY= option sets the transparency for the box
outlines, box fill, whiskers, mean, median, caps, connect lines, and
outliers. You can combine this option with DATATRANSPARENCY=
to set one transparency for the box outlines and the whiskers, mean,
median, caps, and connect lines, but a different transparency for the box
fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

GROUP=column | discrete-attr-var | expression

creates a box plot for each unique group value of the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

The box plot display depends on the setting for the GROUPDISPLAY= option. This
option can be used to group the box plots in the display.

Defaults Each distinct group value is represented in the plot by a different box
outline color. The outline colors are defined by the ContrastColor
attribute of the GraphData1–GraphDataN and GraphMissing style
elements.
BOXPLOTPARM Statement 351

If the MEDIAN, CAPS, or NOTCHES are enabled by the DISPLAY=

option, then each of these features uses the same color as the box
outline. Line styles do not change by group value.

If the MEAN or OUTLIERS are enabled by the DISPLAY= option,

then each distinct group value is represented by a different marker.
The markers are defined by the MarkerSymbol and ContrastColor
attributes of the GraphData1–GraphDataN and GraphMissing style
elements. A marker is used for both MEAN and OUTLIERS, if
displayed.

If box fills are enabled by the ODS style or by the DISPLAY= option,
then each distinct group value is represented in the plot by a different
fill color. The fill colors are defined by the Color attribute of the
GraphData1–GraphDataN and GraphMissing style references.

Interactions Connect lines are not drawn for grouped data.

This option causes the DISPLAY=(CONNECT) and the CONNECT=

options to be ignored.

Tips By default, the group values are mapped in the order of the data. The
GROUPORDER= option can be used to control the sorting order of
the group values. The INDEX= option can be used to alter the default
sequence of colors and markers.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

The representations that are used to identify the groups can be

overridden individually. For example, each distinct group value is
represented by a different line pattern for the box outlines, but the
PATTERN= setting on the OUTLIERATTRS= option could be used to
assign the same line pattern to all box outlines and connect lines.

See the GROUPDISPLAY= option to see the output for the grouped boxes

“DISCRETEATTRVAR Statement” on page 1297

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how to display the boxes that represent group values for the coordinate
pairs. The following example shows a box plot with GROUPDISPLAY=CLUSTER:
352 Chapter 6 • Plot Statements

OVERLAY
draws boxes for a given group value at the exact coordinate. Depending on the
data, boxes at a given coordinate might overlap.
CLUSTER
draws boxes for a given group value adjacent to each other. This option is
available only when the category (X) column is discrete.

Default OVERLAY

Interactions This option is ignored unless GROUP= is specified.

The groups in the legend are shown in the order that is specified in
GROUPORDER by default.

Attributes such as color, symbol, and pattern are assigned to each

group in DATA order by default.

Tip Use the CLUSTERWIDTH= option to control the width of the clusters
when CLUSTER is in effect.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
DESCENDING
orders the groups within a category in descending order.
BOXPLOTPARM Statement 353

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping box attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.
354 Chapter 6 • Plot Statements

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

INTERVALBOXWIDTH=AUTO | dimension
specifies the box width when an interval category (X) column is specified.
AUTO
prior to the third maintenance release of SAS 9.4, AUTO uses 85% of the
smallest interval between any two boxes for the given plot. Starting with the third
maintenance release of SAS 9.4, AUTO uses the BOXWIDTH= option setting or
the Y value when STAT=BOXWIDTH and the corresponding Y value is in
range.
dimension
sets the box width to the specified value.

Default AUTO

Restriction The axis type for the category axis (X by default) must be LINEAR,
and the X column must be numeric.

Interactions Starting with the third maintenance release of SAS 9.4, this option is
overridden by the Y value when the STAT= column value is
BOXWIDTH and the corresponding Y value is in range.

Prior to the third maintenance release of SAS 9.4, this option controls
the box width for an interval box plot. Starting with the third
maintenance release of SAS 9.4, this option overrides the
BOXWIDTH= option for an interval box plot.

See “dimension” on page 1340

LABELFAR=TRUE | FALSE
specifies whether all outliers or only far outliers are labeled.
FALSE
applies the labels specified by the DATALABEL= option to both the outliers and
the far outliers.
TRUE
applies the labels specified by the DATALABEL= option to the far outliers.

Default FALSE

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “Statement Summary” on page 326 for information about outliers.

“boolean ” on page 1339 for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
BOXPLOTPARM Statement 355

Default The Y= column label. If a label is not defined, then the Y= column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

MEANATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the marker representing the mean within the box.

Defaults For non-grouped data, GraphBoxMean style element.

For grouped data, the MarkerSymbol, Markersize, and ContrastColor

attributes of the GraphData1–GraphDataN style elements.

Interaction This option is ignored if the DISPLAY= option does not display the
mean.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Marker Options” on page 1350 for available marker-options.

MEDIANATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the line representing the median within the box.

Defaults For non-grouped data, the GraphBoxMedian style element.

For grouped data, the LineStyle and LineThickness attritubes of the

GraphBoxMedian style element, and the ContrastColor attribute of the
GraphData1–GraphDataN style elements.

Interaction This option is ignored if the DISPLAY= option does not display the
median.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and of the boxes.

Default VERTICAL
356 Chapter 6 • Plot Statements

OUTLIERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the markers representing the outliers.

Defaults For non-grouped data, GraphOutlier style element.

For grouped data, the MarkerSymbol, Markersize, and ContrastColor

attributes of the GraphData1–GraphDataN style elements.

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Marker Options” on page 1350 for available marker-options.

OUTLIERTIP=(role-list)
specifies the information to display when the cursor is positioned over an outlier. If
this option is used, then it replaces all of the information that is displayed by default.
(role-list)
an ordered, space-separated list of unique BOXPLOTPARM roles and user-
defined roles. BOXPLOTPARM roles for OUTLIERTIP include X, Y, STAT, and
DATALABEL.

Note In the data tip, the STAT role displays the text “outlier” or “far
outlier” as applicable.

Tip User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns that are
assigned to the X and Y roles, and also the data column Obs, which is
not assigned to any pre-defined BOXPLOTPARM role. The Obs
column must first be assigned a role:
ROLENAME=(TIP1=OBS)
OUTLIERTIP=(X Y TIP1)

Default The columns assigned to these roles are automatically included in the
data tip information: X and Y .

Requirement To generate data tips, you must include an ODS GRAPHICS ON

statement that has the IMAGEMAP option specified, and write the
graphs to the ODS HTML destination.

Interaction The labels and formats for the OUTLIERTIP variables can be
controlled with the TIPLABEL= and TIPFORMAT= options.

See the ROLENAME= option for specifying user-defined roles.

the TIP= option for specifying the information to display when the
cursor is positioned over a box or whisker in the box plot.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the box outline.

Defaults For non-grouped data, the ContrastColor, LineThickness, and LineStyle

attributes of the GraphOutlines style element.
BOXPLOTPARM Statement 357

For grouped data and filled boxes, the LineStyle and LineThickness
attributes of the GraphOutlines style element, and the ContrastColor
attribute of the GraphData1–GraphDataN style elements.

For grouped data and unfilled boxes, the LineThickness attribute of the
GraphOutlines style element, and the ContrastColor and LineStyle
attributes of the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the OUTLIERTIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles DATAMAX, DATAMIN, MAX, MIN, MEAN,
MEDIAN, N, Q1, Q2, STD, SUMWGT, and X.

Note The roles DATAMAX, DATAMIN, and SUMWGT are valid in the
first maintenance release of SAS 9.4 and later releases.

SPREAD=TRUE | FALSE
specifies whether outliers with the same value are spread out to avoid overlap. For
vertical box plots this means offsetting the outliers horizontally. If this option is
358 Chapter 6 • Plot Statements

false, then outliers with the same value are plotted in the same position, which means
only one is visible.

Default FALSE

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “boolean ” on page 1339 for other Boolean values that you can use.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a box or
whisker in the box plot. If this option is used, then it replaces all of the information
that is displayed by default. Roles for columns that do not contribute to the box plot
can be specified along with roles that do contribute.
(role-list)
an ordered, space-separated list of unique BOXPLOTPARM roles.
BOXPLOTPARM roles for TIP include DATAMAX, DATAMIN, MAX, MIN,
MEAN, MEDIAN, N, Q1, Q2, STD, SUMWGT, and X.

Note The roles DATAMAX, DATAMIN, and SUMWGT are valid in the
first maintenance release of SAS 9.4 and later releases.

Tip Statistics such as N, MIN, and MAX are special roles, They are not
column-based like the X role.

Example The following example displays data tips only for the columns that
are assigned to the roles X (CATEGORY) and the statistic MEAN:
TIP=(X MEAN)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: DATAMAX, DATAMIN, MAX, MIN, MEAN,
MEDIAN, N, Q1, Q2, STD, SUMWGT, and X.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

See the OUTLIERTIP= option for specifying the information to display

when the cursor is positioned over an outlier.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
BOXPLOTPARM Statement 359

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the OUTLIERTIP= or TIP= options are
used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the OUTLIERTIP= or TIP= options are
used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page that is displayed when a box or an outlier marker is
selected.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
box or outlier marker that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

The character column can specify different target URLs for each box and outlier
marker.

Requirements The target URLs for the boxes must be specified in the Q1 statistic
observations, and the target URLs for the outlier markers must be
specified in the OUTLIER statistic observations. URLs that are
specified in observations other than Q1 and OUTLIER are ignored.
360 Chapter 6 • Plot Statements

To generate selectable boxes and outlier markers, you must include

an ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The URL value can be blank for a box or outlier, meaning that no
action is taken when that box or outlier marker is selected.

Example The following vehicle mileage data sample shows box and outlier
URLs specified in column URL for Sedan.
STAT X VALUE DATALABEL URL
...
N Sedan 262.0
MEAN Sedan 08
MEDIAN Sedan 20.00
Q1 Sedan 18.00 ./mileageSedan.html
Q3 Sedan 24.00
STD Sedan 4.23
OUTLIER Sedan 36.00 Honda ./mileageHonda.html
OUTLIER Sedan 35.00 Toyota ./mileageToyota.html
OUTLIER Sedan 35.00 Toyota ./mileageToyota.html
OUTLIER Sedan 38.00 Volkswagen ./mileageVolkswagen.html
MIN Sedan 12.00 Volvo
MAX Sedan 33.00 Volvo
...

WHISKERATTRS=style-element | style-element (line-options) | (line-options)

specifies the line properties of the whiskers and caps.

Defaults For non-grouped data, the GraphBoxWhisker style element.

For grouped data, the LineStyle and LineThickness attritubes of the

GraphBoxWhisker style element, and the ContrastColor attribute of the
GraphData1–GraphDataN style elements.

Restriction The caps are always drawn with a solid line.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interactions This option is ignored if the X= argument is not specified.

BOXPLOTPARM Statement 361

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details

Statement Summary
The BOXPLOTPARM statement requires precomputed input data. One reason to choose
this statement over the BOXPLOT statement is that you can control the computational
technique used to compute various statistics for the box plot, such as the mean, quartiles,
location of fences, outlier definition, and so on. See Appendix 5, “Generalized Macro for
BOXPLOTPARM Data,” on page 1357 for examples of such computations using PROC
SUMMARY and multiple DATA steps.
The BOXPLOTPARM statement displays a single box if given just Y and a STAT
argument. It displays multiple boxes if given both Y and X and a STAT argument and X
has more than one unique value.
By default for numeric or character columns, the category (X) axis is
TYPE=DISCRETE. You can override the default and specify TYPE=LINEAR in the
parent layout, provided that the X column is numeric. The axis for the analysis (Y)
column is always LINEAR. When the X axis is LINEAR, you can use the
INTERVALBOXWIDTH= option to specify the box width.
When ORIENT= VERTICAL, the X (or X2) axis is used for the X column and the Y (or
Y2) axis is used for the Y column. When ORIENT=HORIZONTAL, the X (or X2) axis
is used for the Y column and the Y (or Y2) axis is used for the X column.
Two basic box plot representations can be drawn with the BOXPLOTPARM statement: a
schematic (Tukey) box plot and a skeletal box plot. See the EXTREME= option for
details.
The following figure illustrates the box plot elements:
362 Chapter 6 • Plot Statements

As shown in the figure, the bottom and top edges of the box are located at the 1st
quartile (25th percentile) and 3rd quartile (75th percentile) of the sample. Within the
box, you can display the median (50th percentile) as a line and the mean as a marker (see
the DISPLAY= option).
You can also display markers and data labels for outliers. Outliers are observations that
are more extreme than the upper and lower fences (±1.5 IQR). Outliers that are beyond
upper and lower far fences (±3 IQR) are called FAR OUTLIERS and can also be
identified and labeled. From a graphical perspective, the location of fences along the axis
are known, but there is no line or marker that displays a fence. (See DISPLAY= ,
LABELFAR= , and DATALABEL= options).
Finally, you can control the range represented by the whiskers. By default, the whiskers
are drawn from the upper edge of the box to the MAX value, and from the lower edge of
the box to the MIN value. (See the EXTREME= option.)

Input Data Requirements for the BOXPLOTPARM Statement

At a minimum, valid data for the BOXPLOTPARM statement must provide a numeric
column (Y=) that contains calculated statistics for an analysis, and a string column
(STAT=) that identifies each statistic. The Y column must contain nonmissing values for
the Q1 (25th percentile) and Q3 (75th percentile) statistics. If Y values are missing or
not supplied for other statistic values, then those statistics are not displayed in the plot,
regardless of syntax requests to display them.
For example, a petroleum company uses a turbine to heat water into steam that is
pumped into the ground to make oil more viscous and easier to extract. This process
occurs 20 times daily, and the amount of power (in kilowatts) used to heat the water to
the desired temperature is recorded. The following data show the statistics that are
calculated for one day of this process:
BOXPLOTPARM Statement 363

PowerOutputs Statistic

3180.00 MIN

3340.00 Q1

3487.40 MEAN

3490.00 MEDIAN

3610.00 Q3

4050.00 MAX

20.00 N

To plot the data from the preceding table, the following BOXPLOTPARM statement
uses the Y= and STAT= arguments to generate a single box plot for the recorded
statistics:

BOXPLOTPARM Y=PowerOutputs STAT=Statistic;

If the data contain statistics for multiple days of the process, then a third column in the
data must be present to identify the days that the statistics were recorded. For example,
the following data show the statistics that are calculated for two days of this process:

Day PowerOutputs Statistic

04JUL 3180.00 MIN

04JUL 3340.00 Q1

04JUL 3487.40 MEAN

04JUL 3490.00 MEDIAN

364 Chapter 6 • Plot Statements

Day PowerOutputs Statistic

04JUL 3610.00 Q3

04JUL 4050.00 MAX

04JUL 20.00 N

05JUL 3179.00 MIN

05JUL 3333.50 Q1

05JUL 3471.65 MEAN

05JUL 3419.50 MEDIAN

05JUL 3605.00 Q3

05JUL 3849.00 MAX

05JUL 20.00 N

To plot the data from the preceding table, the BOXPLOTPARM statement needs the Y=,
STAT=, and X= arguments to generate a separate box plot for each day that the statistics
were recorded:

BOXPLOTPARM Y=PowerOutputs STAT=Statistic X=Day;

See Appendix 5, “Generalized Macro for BOXPLOTPARM Data,” on page 1357 for a
more complete example of providing input data for BOXPLOTPARM.

About the DISCRETEOFFSET= Option

The DISCRETEOFFSET= option is useful for graphing multiple response variables side
by side on a common axis. By default within an overlay-type layout, if multiple
BOXPLOTPARM statements are used with different analysis variables, then the boxes
for matching X values are centered on the ticks. Depending on the data, the boxes might
be superimposed. The following code fragment shows the default box positioning:
BOXPLOTPARM Statement 365

layout overlay / cycleattrs=true

yaxisopts=(label="Miles Per Gallon");

boxplotparm x=type y=mpg_city stat=y_stat / name="City" ;

boxplotparm x=type y=mpg_highway stat=y_stat / name="Highway" ;

discretelegend "City" "Highway";

endlayout;

To place the different response values side by side, you can assign a different offset to
each BOXPLOTPARM statement. The BOXWIDTH= option can be used in conjunction
with the DISCRETEOFFSET= option to create narrower boxes when desired.
layout overlay / cycleattrs=true
yaxisopts=(label="Miles Per Gallon");

boxplotparm x=type y=mpg_city stat=y_stat / name="City"

discreteoffset=0.2 ;
boxplotparm x=type y=mpg_highway stat=y_stat / name="Highway"
discreteoffset=-0.2 ;

discretelegend "City" "Highway";

endlayout;
366 Chapter 6 • Plot Statements

Changing Box Plot Display

The SAS System defines graphical style elements that control the display of box plots.
Using these style elements as a starting point, you can change the style attribute values
to achieve a very different appearance for your box plots. For more information, see
“Changing Box Plot Display” on page 330.

Example: BOXPLOTPARM Statement

The following graph was generated by the “Example Program” on page 367:
BUBBLEPLOT Statement 367

Example Program
The following input data generates the box for Sedan in the graph. See Appendix 5,
“Generalized Macro for BOXPLOTPARM Data,” on page 1357 to see the code for
creating all of the data.

STAT X VALUE DATALABEL

...
N Sedan 262
MEAN Sedan 21.0840
MEDIAN Sedan 20
Q1 Sedan 18
Q3 Sedan 24
STD Sedan 4.2346
OUTLIER Sedan 36 Honda
OUTLIER Sedan 35 Toyota
OUTLIER Sedan 35 Toyota
OUTLIER Sedan 38 Volkswagen
MIN Sedan 12
MAX Sedan 33
...

Here is the code for this example.

/* Define the template for the plot. */
proc template;
define statgraph boxplotparm1;
begingraph;
entrytitle "City Mileage for Vehicle Types";
layout overlay;
boxplotparm y=value x=x stat=stat /
datalabel=datalabel spread=true ;
endlayout;
endgraph;
end;
run;

/* Use the BOXCOMPUTE macro to generate the data for this plot. */
%boxcompute(indsn=sashelp.cars,x=type,y=mpg_city,datalabel=make);

/* Generate the plot. */

proc sgrender data=boxdata template=boxplotparm1;
run;

BUBBLEPLOT Statement
Creates a bubble plot of the input data. The locations of the bubble centers correspond to the values of X
and Y columns in the data, and the bubble radii correspond to the values of a SIZE column.
Tip: Starting with the third maintenance release of SAS 9.4, you can use subpixel
rendering with this statement. It is enabled by default. To disable subpixel rendering,
specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 33. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.
368 Chapter 6 • Plot Statements

Syntax
BUBBLEPLOT X=column | expression
Y=column | expression
SIZE=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
BUBBLERADIUSMAX=dimension
specifies the drawing size of the largest bubble.
BUBBLERADIUSMIN=dimension
specifies the drawing size of the smallest bubble.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the numeric column or range attribute map variableto use to
determine the bubble colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bubbles.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bubble fills and bubble
outlines
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display outlined bubbles, filled bubbles, or outlined and
filled bubbles.
DRAWORDER=SIZE | DATA
specifies whether the bubbles are drawn according to bubble size or
according to data order.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled bubble areas.
INDEX=positive-integer-column | expression
specifies indices for mapping bubble attributes (fill and outline) to one of the
GraphData1–GraphDataN style elements.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bubble outlines.
RELATIVESCALE=TRUE | FALSE
specifies whether the SIZE= column values are interpreted as relative values.
RELATIVESCALETYPE=LINEAR | PROPORTIONAL
specifies the type of scaling that is to be applied to the SIZE= column values.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.
SIZETHRESHOLDMAX=numeric-value
specifies a SIZE= column value threshold at which bubble size is clamped to
the BUBBLERADIUSMAX= option value.

Axes options
PRIMARY=TRUE | FALSE
BUBBLEPLOT Statement 369

specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
bubbles.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
specifies a column for bubble labels.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the bubble labels.
DATALABELPOSITION=TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |
RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the bubble labels relative to the bubble.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate bubble color for each unique grouping that is specified.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a bubble is selected.
370 Chapter 6 • Plot Statements

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies the column for the X values of the bubble centers.
Y=column | expression
specifies the column for the Y values of the bubble centers.
SIZE=numeric-column | expression
specifies the bubble SIZE values.

Optional Arguments
BUBBLERADIUSMAX=dimension
specifies the drawing size of the largest bubble.

Default Three times as large as the size set by GraphDataDefault:markerSize

Restriction The dimension value must be greater than the

BUBBLERADIUSMIN=dimension value.

Interaction This option is ignored when RELATIVESCALE= FALSE.

Tip A maximum size that is specified as a percent is interpreted as a

percent of the graph's height. The height can be adjusted with the
DESIGNHEIGHT= option in the BEGINGRAPH statement or the
HEIGHT= option in the ODS GRAPHICS statement.

See “dimension” on page 1340

BUBBLERADIUSMIN=dimension
specifies the drawing size of the smallest bubble.

Default GraphDataDefault:markerSize

Restriction The BUBBLERADIUSMIN= value must be less than the

BUBBLERADIUSMAX= value.

Interaction This option is ignored when RELATIVESCALE= FALSE.

Tip A maximum size that is specified as a percent is interpreted as a

percent of the graph's height. The height can be adjusted with the
DESIGNHEIGHT= option in the BEGINGRAPH statement or the
HEIGHT= option in the ODS GRAPHICS statement.

See “dimension” on page 1340

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:
BUBBLEPLOT Statement 371

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorRamp style element for filled bubbles and

ThreeColorAltRamp for unfilled bubbles

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to the
color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the numeric column or range attribute map variableto use to determine the
bubble colors.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Interactions The mapped color used for the bubbles is also used for the data labels.

When the GROUP= option is specified with the COLORRESPONSE=

option, the GROUP= option is ignored.

When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map.

When only the outlines are displayed, this option overrides suboption
COLOR= in the OUTLINEATTRS= option and varies the outline
color according to the color gradient or the attribute map.
372 Chapter 6 • Plot Statements

Note When both the fill and outline are displayed, the bubble fill color
varies according to the color gradient or attribute map but the bubble
outline color remains fixed on the color specified in option
OUTLINEATTRS=.

Tip To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

DATALABEL=column | expression
specifies a column for bubble labels. The label positions are adjusted to prevent them
from overlapping.

Default No labels are displayed.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the bubble labels.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the text color is derived from the GraphData1–
GraphDataN style elements. The data label color changes to match the
group color derived from the ContrastColor attribute of the style
element that is in effect.

Interaction The default attributes are overridden if the COLORRESPONSE=

option is used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

DATALABELPOSITION=TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |

RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the bubble labels relative to the bubble.

Default TOPRIGHT

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters. When set to
TRUE, the data label is split unconditionally at each occurrence of any of the
specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
BUBBLEPLOT Statement 373

data label. In that case, all of the specified split characters together are treated as a
single split character.
When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is
split unconditionally at each occurrence of any of the specified split characters. If the
data label does not contain any of the specified characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
TRUE
drops the split characters from the data label.
FALSE
includes the split characters in the data label. When DATALABELSPLIT=TRUE
and DATALABELSPLITCHARDROP=FALSE, each split character remains as
the last character in the current line. The characters that follow the split character,
up to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a data label with the following
specifications:
• the data label text for this label is Product*Group*A
• DATALABELSPLIT=TRUE
• DATALABELSPLITCHARDROP=TRUE | FALSE
• DATALABELSPLITCHAR="*"
374 Chapter 6 • Plot Statements

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from the

label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains as the
last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.
AUTO
justifies the labels based on the DATALABELPOSITION= option as shown in
the following table.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which DATALABELPOSITION=TOP.
Note: The gray vertical line at the bottom of each label represents the horizontal
center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of text.

The text box is anchored the same way that the unsplit text is anchored. For example,
if DATALABELPOSITION=TOP, then the bottom center of the text box is
positioned at the top of the marker.
BUBBLEPLOT Statement 375

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bubbles. The following figure shows
bubbles with each of the skins applied.

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, DISPLAY= FILL must be in effect.
Otherwise, this option is ignored.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The appearance of the data skin is based on the FILLATTRS= color.

This option is ignored if the RELATIVESCALE= option is set to

FALSE.

When a data skin is applied, all bubble outlines are set by the skin,
and the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bubble fills and bubble outlines

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

376 Chapter 6 • Plot Statements

Note This option does not affect the data labels.

Tip The FILLATTRS= option can be used to set transparency for just the filled
bubble areas. You can combine this option with FILLATTRS= to set one
transparency for the bubble outlines but a different transparency for the
bubble fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display outlined bubbles, filled bubbles, or outlined and filled
bubbles.
STANDARD
displays outlined, filled bubbles
ALL
displays outlined, filled bubbles
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays outlined bubbles

FILL displays filled bubbles

Default STANDARD

Tip Use the DATASKIN= , OUTLINEATTRS= , and FILLATTRS= options to

control the appearance of the bubbles.

DRAWORDER=SIZE | DATA
specifies whether the bubbles are drawn according to bubble size or according to
data order.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
SIZE
draws the bubbles according to bubble size, from the largest to the smallest
DATA
draws the bubbles according to data order
The following figure shows the effect of SIZE and DATA on four bubbles. The
bubble labels indicate the data order, and the bubble sizes increase linearly starting
with 1.

Default SIZE
BUBBLEPLOT Statement 377

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled bubble areas.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

For grouped data, the Color attribute of GraphData1–GraphDataN

style elements.

Interactions For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is
ignored, and the bubble fill colors vary according to the gradient.

Tip The DATATRANSPARENCY= option sets the transparency for the

bubble fills and the bubble outlines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines but
a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

GROUP=column | discrete-attr-var | expression

creates a separate bubble color for each unique grouping that is specified.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default The bubble attributes for each unique group value are derived from the
GraphData1–GraphDataN and GraphMissing style elements. If the
bubbles are filled, then the COLOR attribute is used for bubble fill and
the CONTRASTCOLOR attribute is used for the bubble outlines. If
the bubbles are not filled, then the CONTRASTCOLOR and
PATTERN attributes are used for the bubble outlines.

Interactions If a discrete attribute map variable is specified, then the color mapping
for the bubbles is defined by the associated DISCRETEATTRMAP
statement. See “DISCRETEATTRMAP Statement” on page 1287.

The mapped color that is used for outlines is also used as the color of
the data labels.

This option is ignored if the COLORRESPONSE= option is also used.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.
378 Chapter 6 • Plot Statements

Note All bubbles have only one fill and one outline color as specified by the
FILLATTRS= and OUTLINEATTRS options.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping bubble attributes (fill and outline) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The SIZE= column label. If a label is not defined, then the SIZE=
column name is used.
BUBBLEPLOT Statement 379

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bubble outlines.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, unfilled bubbles use both the CONTRASTCOLOR

and PATTERN attributes of the GraphData1–GraphDataN style
elements. Filled bubbles use only the CONTRASTCOLOR attribute.
If the COLORRESPONSE= option is specified and the bubbles are
filled, then the outline attributes are derived from the
GraphDataDefault style element. For unfilled bubbles, the outline
colors vary according to the gradient.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the bubble outline colors vary according to the gradient.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
380 Chapter 6 • Plot Statements

multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

RELATIVESCALE=TRUE | FALSE
specifies whether the SIZE= column values are interpreted as relative values.
Relative means that the size values do not translate directly into bubble radii. Rather,
the bubble sizes are scaled to represent the value range of the SIZE= column.
For example, when RELATIVESCALE=TRUE, if only two bubbles are drawn with
sizes of 2 and 4, then they appear the same as only two bubbles with sizes of 4000
and 8000. By contrast, when RELATIVESCALE=FALSE, the size values are
interpreted in the same units as the axes.
If you set this option to FALSE, then it is recommended that you also place the
BUBBLEPLOT statement in a LAYOUT OVERLAYEQUATED container. This
ensures that the X and Y axis units are the same. For more information, see “Details”
on page 384.

Default TRUE

Interactions When this option is set to TRUE, the BUBBLERADIUSMAX= and

BUBBLERADIUSMIN= options can be used to fix the drawing size
of the smallest bubble and largest bubble. If
RELATIVESCALE=FALSE, then the BUBBLERADIUSMAX= and
BUBBLERADIUSMIN= options are ignored.

If this option is set to FALSE, then the DATASKIN= option is

ignored.

If one or both axes are discrete, then RELATIVESCALE=FALSE is

ignored.

Tip If you specify RELATIVESCALE=FALSE, then it is recommended

that you also place the BUBBLEPLOT statement in a LAYOUT
OVERLAYEQUATED container to ensure that the X and Y axis units
are the same. If you place the BUBBLEPLOT statement in a
LAYOUT OVERLAY container instead, then the bubbles might be
drawn as ellipses because the X and Y axes units are different.

See “boolean ” on page 1339 for other Boolean values that you can use.

RELATIVESCALETYPE=LINEAR | PROPORTIONAL
specifies the type of scaling that is to be applied to the SIZE= column values.
LINEAR
increases the size of the bubbles in linear proportion to the range of the SIZE=
column values. For example, if only two bubbles are drawn with sizes of 2 and 4,
then they appear the same as only two bubbles with sizes of 4000 and 8000.
PROPORTIONAL
increases the size of each bubble in direct proportion to its corresponding SIZE=
column value. For example, if only two bubbles are drawn with sizes of 50 and
100, then the bubble for SIZE=50 is drawn to half the size of the bubble for
SIZE=100.
BUBBLEPLOT Statement 381

Default LINEAR

Interactions This option is ignored when RELATIVESCALE=FALSE.

When theSIZETHRESHOLDMAX= option is specified, for any

SIZE= column value that is greater than the
SIZETHRESHOLDMAX= value, the proportional scale is adjusted so
that the size of the bubble for that value is clamped to
theBUBBLERADIUSMAX= value.

If all the values for the SIZE= column are negative, then
RELATIVESCALETYPE=PROPORTIONAL is ignored, and the
default value is used.

When RELATIVESCALETYPE=PROPORTIONAL is specified, the

BUBBLERADIUSMIN= option specifies the minimum bubble size.
In that case, when a SIZE= column value results in a bubble of a size
that is less than the BUBBLERADIUSMIN= value, the bubble size for
that value is changed to the BUBBLERADIUSMIN= value.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either the
ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y , SIZE, GROUP, DATALABEL, and
COLORRESPONSE.

SIZETHRESHOLDMAX=numeric-value
specifies a SIZE= column value threshold at which bubble size is clamped to the
BUBBLERADIUSMAX= option value. The size of the bubbles for all SIZE=
column values that equal or exceed the specified threshold value is set to the
BUBBLERADIUSMAX= value.

Default The maximum SIZE= column value is mapped to the

BUBBLERADIUSMAX= option value.
382 Chapter 6 • Plot Statements

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the bubbles. If
this option is used, then it replaces all of the information that is displayed by default.
Roles for columns that do not contribute to the bubble plot can be specified along
with roles that do.
(role-list)
an ordered, space-separated list of unique BUBBLEPLOT and user-defined roles.
BUBBLEPLOT roles include X, Y, SIZE, GROUP, DATALABEL, and
COLORRESPONSE.

Tip User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns assigned to
the roles X, Y, and SIZE, as well as the column Pop_2009. The
POP_2009 column is not assigned to any pre-defined BUBBLEPLOT
role, so it must first be assigned a role:

ROLENAME=(TIP1=POP_2009)
TIP=(TIP1 X Y SIZE)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X , Y , SIZE , GROUP , DATALABEL , and
COLORRESPONSE .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.
BUBBLEPLOT Statement 383

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a bubble is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
bubble that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate a plot with selectable bubbles, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding point is selected.

The URL value can be the same for any X and Y pairs. In that case,
the same action is taken when the bubbles for those X and Y pairs are
selected.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X
384 Chapter 6 • Plot Statements

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
The BUBBLEPLOT statement displays one bubble for each row in the data, provided
that row contains nonmissing values for X, Y, and SIZE. By default, the bubbles are
displayed as filled, outlined circles. Regardless of the data order, the bubbles are always
drawn from the largest size to the smallest size.
By default, the minimum and maximum values of the SIZE= column establish a range
over which the bubble radii increase in linear proportion. The actual drawing size of the
smallest and largest bubble is set automatically. You can adjust the smallest and largest
bubble sizes with the BUBBLERADIUSMIN= and BUBBLERADIUSMAX= options.
In these cases where the bubble sizes are proportional to each other, the default setting
RELATIVESCALE=TRUE is appropriate.
If the SIZE= values are in the same units as the X and Y values, and both X and Y are
numeric, then you can generate a plot where the bubble-radius units match the axis-scale
units. To do so, specify the BUBBLEPLOT statement within a LAYOUT
OVERLAYEQUATED block, and in the BUBBLEPLOT statement, set
RELATIVESCALE=FALSE.
By default, for character columns, the X and Y axes are always discrete. For numeric
columns, the X and Y axes are linear. You can change axis type for numeric axes with
the layout options XAXISOPTS= and YAXISOPTS=.
Note: Within a LAYOUT OVERLAY, the unit-interval of the X and Y axes are not
necessarily the same and the bubbles might be distorted into ellipses when
RELATIVESCALE=FALSE. The OVERLAYEQUATED container ensures that the
bubbles are displayed as circles, assuming that both the X= and Y= arguments
specify numeric columns.
data influence;
input x y radius category;
datalines;
2 4 1 1
5 5 2 1
6 3 2 2
12 7 3 2
;
proc template;
define statgraph equatedbubbles;
begingraph;
entrytitle 'Radius of Influence';
entrytitle 'Bubbles Show Distance Covered by Observation';
layout overlayequated /
Example: BUBBLEPLOT Statement 385

xaxisopts = (griddisplay=on)
yaxisopts = (griddisplay=on);
bubbleplot x=x y=y size=radius /
group=category datatransparency=0.5
relativescale=false ;
endlayout;
endgraph;
end;

proc sgrender data=influence template=equatedbubbles;

run;

Example: BUBBLEPLOT Statement

The following graph was generated by the “Example Program” on page 386 :
386 Chapter 6 • Plot Statements

Example Program
data bubbleintro;
input Engineer $ Salary number;
format Salary dollar7.0 number comma6.0;
datalines;
Electric 59000 89382
Civil 54000 73273
Software 56000 34833
Chemical 62000 25541
Mechanical 60000 19601
;
proc template;
define statgraph engineer;
begingraph;
entrytitle 'Median Salary for Entry Level Engineers';
entrytitle 'Bubbles Show Number of Engineers in Survey';
layout overlay;
bubbleplot x=engineer y=salary
size=number / datalabel=number;
endlayout;
endgraph;
end;

proc sgrender data=bubbleintro template=engineer;

run;

CONTOURPLOTPARM Statement
Creates a contour plot representing a response variable evaluated over a grid of X and Y values.
Restriction: Contour plots do not support data tips.
CONTOURPLOTPARM Statement 387

Tips: By default, the CONTOURPLOTPARM statement assumes that the X-Y grid is
complete and does not contain any missing or irregular values. If the X-Y grid is not
complete, specify GRIDDED=FALSE in the CONTOURPLOTPARM statement so
that the values needed to complete the grid are calculated. Otherwise, unexpected
results might occur.
You can use a legend to display the contour level values. For
CONTOURTYPE=LINE and CONTOURTYPE=LABELEDLINE, use a
DISCRETELEGEND statement to add a legend. For all other contour types, use a
CONTINUOUSLEGEND statement to add a legend.
For filled contour types, there might be small, visible gaps between the axes and the
contour boundaries. To eliminate the gaps, specify the following axis options in the
layout statement for the plots parent layout:
XAXISOPTS=(OFFSETMIN=0 OFFSETMAX=0
LINEAROPTS=(THRESHOLDMIN=0 THRESHOLDMAX=0))
YAXISOPTS=(OFFSETMIN=0 OFFSETMAX=0
LINEAROPTS=(THRESHOLDMIN=0 THRESHOLDMAX=0))

Starting with the third maintenance release of SAS 9.4, you can use subpixel
rendering with this statement. It is enabled by default. To disable subpixel rendering,
specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 33. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.

Syntax
CONTOURPLOTPARM X=numeric-column | expression
Y=numeric-column | expression
Z=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used to determine the colors of filled or
gradient contours.
CONTOURTYPE=LINE | FILL | GRADIENT | LINEFILL | LINEGRADIENT |
LABELEDLINE | LABELEDLINEFILL | LABELEDLINEGRADIENT
specifies how the contour is displayed.
GRIDDED=TRUE | FALSE
specifies whether the X and Y values are equally spaced in a rectangular grid.
LEVELS=(contour-value-list)
specifies a list of contour level values.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the contour lines.
NHINT=integer
specifies the suggested number of contour levels for the Z column.
NLEVELS=integer
specifies the actual number of contour levels for the Z column.
REVERSECOLORMODEL=TRUE | FALSE
388 Chapter 6 • Plot Statements

specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
LINELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the contour line labels.
LINELABELBASELINE=HORIZONTAL | TANGENT
specifies the text alignment of the contour line labels.
LINELABELFORMAT=format
specifies the format to use for the contour line labels.
LINELABELPOSITION=MIDDLE | BEGIN | END
specifies the position for the contour line labels.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the X coordinates for the grid.
Y=numeric-column | expression
specifies the Y coordinates for the grid.
Z=numeric-column | expression
specifies the contour response values.

Optional Arguments
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used to determine the colors of filled or gradient
contours.
style-element
specifies the name of a style element. The style element can contain these style
attributes:

STARTCOLOR specifies a color for the smallest data value of the Z=

column.
CONTOURPLOTPARM Statement 389

NEUTRALCOLOR specifies a color for the midpoint of the range of the

Z= column.
ENDCOLOR specifies a color for the highest data value of the Z=
column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData2:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorRamp style element.

Interactions The REVERSECOLORMODEL= option can be used to reverse the

start and end colors of the ramp assigned to the color model.

The NEUTRALCOLOR attribute is not used for two-color ramps.

CONTOURTYPE=LINE | FILL | GRADIENT | LINEFILL | LINEGRADIENT |

LABELEDLINE | LABELEDLINEFILL | LABELEDLINEGRADIENT
specifies how the contour is displayed.
LINE
displays contour levels as unlabeled lines.
FILL
displays the area between the contour levels as filled. Each contour interval is
filled with one color.
GRADIENT
displays a smooth gradient of color to represent contour levels.
LINEFILL
combines the LINE and FILL types. Each contour interval is filled with one
color. Displays contour levels as unlabeled lines.
LINEGRADIENT
combines the LINE and GRADIENT types. Displays contour levels as unlabeled
lines.
LABELEDLINE
adds labels to the LINE type, displaying contour levels as labeled lines.
LABELEDLINEFILL
adds labels to the LINEFILL type. Each contour interval is filled with one color.
Displays contour levels as lines with labels showing contour level values.
LABELEDLINEGRADIENT
adds labels to the LINEGRADIENT type. Displays contour levels as lines with
labels showing contour level values.
The following figure shows the effect of each of the values.
390 Chapter 6 • Plot Statements

Default The GraphContour:DisplayOpts style reference.

Interactions The fill colors of the types that enable FILL or GRADIENT are
controlled by the COLORMODEL= option.

The line properties of the types that enable LINE or LABELEDLINE

are controlled by the LINEATTRS= option.

The label properties of the types that enable LABELEDLINE are

controlled by the LINELABELATTRS= and
LINELABELBASELINE= options.

If a DISCRETELEGEND statement is associated with the contour,

then the legend is NOT displayed if CONTOURTYPE= is set to FILL
or GRADIENT.

If a CONTINUOUSLEGEND statement is associated with the

contour, then the legend is NOT displayed if CONTOURTYPE is set
to LINE or LABELEDLINE.

GRIDDED=TRUE | FALSE
specifies whether the X and Y values are equally spaced in a rectangular grid. If set
to FALSE, then additional calculations are performed in order to complete the grid.
For information about the algorithm used to calculate the grid, see “Mesoscale
Objective Map Analysis Using Weighted Time-Series Observations.”1

Default TRUE

Tip By default, the CONTOURPLOTPARM statement assumes that the X-Y

grid is complete and does not contain any missing or irregular values. If the
X-Y grid is not complete, then specify GRIDDED=FALSE so that the plot
calculates the values needed to complete the grid. Otherwise, unexpected
results might occur.

See “boolean ” on page 1339 for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Barnes,
1 Stanley L March 1973. “Mesoscale Objective Map Analysis Using Weighted Time-Series Observations.” Technical
Memorandum (NOAA TM ERLNSSL-62), United States National Oceanic and Atmospheric Administration, Environmental
Research Labs, Norman, OK..
CONTOURPLOTPARM Statement 391

Default The Z-column label. If a label is not defined, then the Z-column name
is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LEVELS=(contour-value-list)
specifies a list of contour level values.
(contour-value-list)
a space-separated list of numeric values, enclosed in parentheses.

Default The number of levels and the level values are determined internally
using the NHINT= or NLEVELS= option value.

Interaction This option overrides the NHINT= and NLEVELS= options.

Note Values that are outside of the data range are ignored.

Example levels=(0.0001 0.0004 0.0007 0.0010 0.0013 0.0016)

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the contour lines.

Default The GraphDataDefault style element.

Interaction This option is honored only if the CONTOURTYPE= displays lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

LINELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the contour line labels.

Default The GraphValueText style element.

Interaction This option is honored only if the CONTOURTYPE= displays labels.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LINELABELBASELINE=HORIZONTAL | TANGENT
specifies the text alignment of the contour line labels. Each contour line has a
precomputed label point.
HORIZONTAL
specifies that each label is parallel to the X-axis. The label intersects its contour
line and is centered at the label point.
TANGENT
specifies that each label is drawn tangent to the contour line at the label point.
This reduces intersection with the contour line and causes labels to be tilted at
various angles in relation to the X-axis.

Default HORIZONTAL
392 Chapter 6 • Plot Statements

Interaction This option is honored only if the CONTOURTYPE= displays labels.

LINELABELFORMAT=format
specifies the format to use for the contour line labels.

Default The format associated with the Z column or BEST6. if no format is

assigned.

Interaction This option is honored only if the CONTOURTYPE= displays labels.

LINELABELPOSITION=MIDDLE | BEGIN | END

specifies the position for the contour line labels.

Default MIDDLE

Interaction This option is honored only if the CONTOURTYPE= option specifies

labels.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to indicate the colors
associated with the Z values.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

NHINT=integer
specifies the suggested number of contour levels for the Z column.

Default 7

Interaction This option is ignored if the LEVELS= or NLEVELS= option is

specified.

Note The actual number of levels is adjusted to provide an appropriate

number of levels for the data.

NLEVELS=integer
specifies the actual number of contour levels for the Z column.

Default The number of levels is determined internally, using the NHINT=

value.

Interactions This option overrides the NHINT= option.

This option is ignored if CONTOURTYPE= GRADIENT.

This option is ignored if the LEVELS= option is specified.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.
CONTOURPLOTPARM Statement 393

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either the
ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page 1339 for other Boolean values that you can use.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
A CONTOURPLOTPARM statement uses the CONTOURTYPE= option to specify the
type of contour plot to generate. Contour types that display fills or gradients but no
contour lines can use only a CONTINUOUSLEGEND statement to represent the contour
level values in a legend. Contour types that display lines can use either a
CONTINUOUSLEGEND or DISCRETELEGEND statement to identify contour level
values.
By default, the CONTOURPLOTPARM statement assumes that the X-Y grid is
complete. If the grid is not complete, then set the GRIDDED= option to FALSE so that
the plot calculates the values needed to complete into a grid with a bounding rectangle.
394 Chapter 6 • Plot Statements

For filled or gradient contour types, small gaps might be visible between the axes and
the bounding box of the contour data. To eliminate these gaps, set the axis options of the
LAYOUT OVERLAY statement as follows:

XAXISOPTS=(OFFSETMIN=0 OFFSETMAX=0
LINEAROPTS=(THRESHOLDMIN=0 THRESHOLDMAX=0))

YAXISOPTS=(OFFSETMIN=0 OFFSETMAX=0
LINEAROPTS=(THRESHOLDMIN=0 THRESHOLDMAX=0))

Contour plots do not support the data tips that are enabled by the IMAGEMAP= option
in the ODS GRAPHICS statement.

Example: CONTOURPLOTPARM Statement

The following graph was generated by the “Example Program” on page 394:

Example Program
proc template;
define statgraph contourplotparm;
begingraph;
entrytitle "Contour Plot of Height and Weight";
layout overlay /
xaxisopts=(offsetmin=0 offsetmax=0
linearopts=(thresholdmin=0 thresholdmax=0))
yaxisopts=(offsetmin=0 offsetmax=0
linearopts=(viewmax=250
thresholdmin=0 thresholdmax=0));
DENDROGRAM Statement 395

contourplotparm x=height y=weight z=density /

contourtype=fill nhint=12
name="Contour" colormodel=twocolorramp;
continuouslegend "Contour" / title="Density";
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.gridded template=contourplotparm;
run;

DENDROGRAM Statement
Creates a tree diagram that is typically used to display the results of a hierarchical clustering analysis.

Syntax
DENDROGRAM NODEID=column | expression
PARENTID=column | expression
CLUSTERHEIGHT=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
CLUSTERS=numeric-column | expression
specifies a numeric column containing the resultant number of clusters at
each node.
CUT=TRUE | FALSE
specifies whether the tree is to be cut.
CUTOPTS=(pruning-options)
specifies pruning options for cutting the dendrogram.
DATATRANSPARENCY=number
specifies the degree of the transparency of the dendrogram lines.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the dendrogram lines.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the dendrogram leaf axis.
TREETYPE=RECTANGULAR | TRIANGULAR
specifies the type of tree structure to draw.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
396 Chapter 6 • Plot Statements

Data tip options

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
dendrogram line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
NODEID=column | expression
specifies a column for the ID values of the nodes. Each node ID value must be
unique. If duplicate NODEID values are found, then the dendrogram is not
rendered . The column can be numeric or character, but it must be of the same type
and have the same formatted length as the ParentID column.
The maximum number of nodes that are supported by the dendrogram is determined
by the DISCRETEMAX= option in the ODS GRAPHICS statement. The default
value is DISCRETEMAX=1000. If the graph data contains more than 1000 discrete
values, then the dendrogram is not drawn and a warning is written to the SAS log. In
that case, you can use the DISCRETEMAX= option to increase the maximum
number of discrete values that are allowed.
PARENTID=column | expression
specifies a column for the parent ID values of the nodes. The column can be numeric
or character, but it must be of the same type and have the same formatted length as
the NodeID column.
CLUSTERHEIGHT=numeric-column | expression
specifies the column for the height values for each node.

Optional Arguments
CLUSTERS=numeric-column | expression
specifies a numeric column containing the resultant number of clusters at each node.

Interaction For this option to take effect, the pruning options in the CUTOPTS=
option must set TYPE=NCLUSTERS and specify a number for the
NCLUSTERS= setting.

CUT=TRUE | FALSE
specifies whether the tree is to be cut.

Default FALSE

Tip To set the properties of the CUT, use the CUTOPTS= option.
DENDROGRAM Statement 397

See “boolean ” on page 1339 for other Boolean values that you can use.

CUTOPTS=(pruning-options)
specifies pruning options for cutting the dendrogram. The following pruning-options
must be specified as a space-separated list of option = value pairs enclosed in
parentheses.
CUTHEIGHT=number
specifies the height at which the tree is to be pruned.

Default The tree is not pruned.

Requirement You must include pruning option TYPE=CUTHEIGHT with this

option.

Interaction This option is ignored when CUT=FALSE or when pruning

option TYPE=CUTHEIGHT is not explicitly specified.

Example CUTOPTS=(TYPE=CUTHEIGHT CUTHEIGHT=0.75)

NCLUSTERS=number
specifies the number of clusters to use for pruning the tree.

Default The tree is not pruned.

Interaction For this setting to take effect, pruning-option TYPE=NCLUSTERS

must also be set. In addition, the CLUSTERS= option must be used,
and the CUT= option must be set to TRUE.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the cut lines.

Default The GraphDataDefault style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

TYPE=CUTHEIGHT | NCLUSTERS
specifies which rule to use to prune the tree.

Default CUTHEIGHT

DATATRANSPARENCY=number
specifies the degree of the transparency of the dendrogram lines.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The CLUSTERHEIGHT= column label. If a label is not defined, then

the CLUSTERHEIGHT= column name is used.
398 Chapter 6 • Plot Statements

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the dendrogram lines.

Default The GraphDataDefault style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the dendrogram leaf axis.

Default VERTICAL

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a dendrogram
line. If this option is used, then it replaces all of the information that is displayed by
default.
(role-list)
an ordered, space-separated list of unique DENDROGRAM roles, which include
NODEID , PARENTID , and CLUSTERHEIGHT .
DENDROGRAM Statement 399

Example The following example displays data tips for the columns assigned to
the roles NODEID and PARENTID.
TIP=(NODEID PARENTID)

NONE
suppresses data tips from the plot.

Default The columns assigned to the following roles are automatically

included in the data tip information: NODEID , PARENTID , and
CLUSTERHEIGHT .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(CLUSTERHEIGHT=4.1)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(CLUSTERHEIGHT="Height")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

TREETYPE=RECTANGULAR | TRIANGULAR
specifies the type of tree structure to draw.
400 Chapter 6 • Plot Statements

Default RECTANGULAR

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
A dendrogram is a tree diagram that is typically used to show the cluster arrangements in
hierarchical data. The DENDROGRAM statement supports clusters with only a single
root. If multiple roots are found in the data, then a warning is written to the SAS log and
the dendrogram is not drawn. The maximum number of nodes that is supported by the
DENDROGRAM statement is determined by the DISCRETEMAX= option in the ODS
GRAPHICS statement, which is 1000 by default.
In the Graph Template Language, a DENDROGRAM plot typically appears by itself in a
LAYOUT OVERLAY container. You can overlay REFERENCELINE or BANDPLOT
statements on a DENDROGRAM, but overlaying other plot types might produce
unexpected results.
Using the DENDROGRAM statement in layouts where the axis ranges are merged
across cells might produce unexpected results.

Example: DENDROGRAM Statement

The following graph was generated by the “Example Program” on page 401:
Example: DENDROGRAM Statement 401

Example Program
data clustree;
input id $ parent $7-12 height nClus;
label id="Cluster ID" parent="Parent ID";
datalines;
clus1 3 1
clus2 clus1 0.2 7
clus3 clus1 1.75 2
clus4 clus3 0.7 4
clus5 clus3 0.8 3
clus6 clus4 0.4 5
clus7 clus6 0.1 9
clus8 clus5 0.25 6
clus9 clus8 0.15 8
1 clus9 0 10
2 clus6 0 10
3 clus2 0 10
4 clus7 0 10
5 clus7 0 10
6 clus2 0 10
7 clus4 0 10
8 clus5 0 10
9 clus8 0 10
10 clus9 0 10
run;

proc template;
define statgraph dendrogram;
begingraph;
layout overlay;
dendrogram nodeID=id parentID=parent clusterheight=height;
402 Chapter 6 • Plot Statements

endlayout;
endgraph;
end;
run;

proc sgrender data=clustree template=dendrogram;

run;

DENSITYPLOT Statement
Creates a univariate probability density curve computed from input data.
Tips: If the data density is not known, then use the KERNEL distribution option in the
DENSITYPLOT statement.
Starting with the third maintenance release of SAS 9.4, subpixel rendering is enabled
by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page
33.For information about the ODS GRAPHICS statement SUBPIXEL= option, see
“ODS GRAPHICS Statement” in SAS ODS Graphics: Procedures Guide.

Syntax
DENSITYPLOT numeric-column | expression </<distribution-option> <option(s)>>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the density curve and curve label.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the density curve.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.
DENSITYPLOT Statement 403

Label options
CURVELABEL="string"
specifies a label for the density curve.
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the density curve label.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the density curve label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the density curve label relative to the curve line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the curve label block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate density curve for each unique group value in the specified
column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation
that is read.
WEIGHT=numeric-column | expression
specifies a column that contains a density-curve calculation a priori weight
for each observation of the input data object.

Required Arguments
numeric-column
specifies a numeric column of data values that are used to calculate the parameters
for the probability distribution.
expression
specifies an expression that calculates values when those values are not stored in the
data.
404 Chapter 6 • Plot Statements

Optional Arguments
CURVELABEL="string"
specifies a label for the density curve.

Default No curve label is displayed.

Restriction This option is not valid when the GROUP= option is specified.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the density curve label.

Default The GraphValueText style element.

Interaction For this option to take effect, the CURVELABEL=curvelabel option

must also be used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the density curve label relative to the plot area.
INSIDE
locates the labels inside the plot area
OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the DENSITYPLOT is used in multi-

cell layouts such as LATTICE, DATAPANEL, or DATALATTICE,
where axes can be external to the grid.

Interaction This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the curve label
appears. For more information, see “Location and Position of Curve
Labels” on page 185.

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the density curve label relative to the curve line.
AUTO
positions the density label automatically near the end of the density curve along
unused axes whenever possible (typically Y2 or X2) to avoid collision with tick
values.

Restriction This option is used only when

CURVELABELLOCATION=OUTSIDE.

MAX
forces the density label to appear near maximum density X-values (typically, to
the right).
DENSITYPLOT Statement 405

MIN
forces the density label to appear near minimum density X-values (typically, to
the left).
START
forces the density label to appear near the beginning of the curve.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

END
forces the density label to appear near the end of the curve.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION=INSIDE

is specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the density
label appears. For more information, see “Location and Position of
Curve Labels” on page 185.

Note When you specify TICKVALUELIST=, VIEWMAX=, or

VIEWMIN= in an axis statement, the data points that are used to
determine the position of the curve label might fall outside of the
graph area. In that case, the curve label might not be displayed or
might be positioned incorrectly.

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters. When a
curve label is split, the label is split on each occurrence of the specified split
characters.

Default FALSE. The curve label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

406 Chapter 6 • Plot Statements

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
curve label. In that case, all of the specified split characters together are treated as a
single split character.
When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the curve
label is split unconditionally at each occurrence of any of the specified split
characters. If the curve label does not contain any of the specified characters, then
the label is not split.
"character-list"
one or more characters with no delimiter between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the following
option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the curve label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the curve label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.
TRUE
drops the split characters from the curve label text.
FALSE
includes the split characters in the curve label text. When
CURVELABELSPLIT=TRUE and
CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split character, up
to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a curve label with the following
specifications:
DENSITYPLOT Statement 407

• CURVELABELPOSITION=MAX
• CURVELABEL="Product*Group*A"
• CURVELABELSPLIT=TRUE
• CURVELABELSPLITCHARDROP=TRUE | FALSE
• CURVELABELSPLITCHAR="*"
Note: The horizontal line to the left of the label represents the maximum end of the
curve for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the curve label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the curve label block.
AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown in
the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which CURVELABELPOSITION=MAX.
Note: The horizontal line to the left of each label represents the maximum end of the
curve for reference.
408 Chapter 6 • Plot Statements

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the density curve and curve label.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that is
read.

Default All observations have a frequency count of 1.

Restriction If the value of the numeric-column is missing or is less than 1, then the
observation is not used in the analysis. If the value is not an integer,
then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation.

GROUP=column | discrete-attr-var | expression

creates a separate density curve for each unique group value in the specified column.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

For example, the Sashelp.Cars data contains a column named Origin, which
identifies the region that produces each car. This column could be used in the
DENSITYPLOT statement to group the density curves in the display:
proc template;
define statgraph densityplot;
begingraph;
entrytitle "Highway Mileage Distribution by Origin";
DENSITYPLOT Statement 409

layout overlay /
xaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot))
yaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot));
densityplot mpg_highway / name="densityplot" group=origin;
discretelegend "densityplot" / title="Origin:";
endlayout;
endgraph;
end;
run;

proc sgrender template=densityplot data=sashelp.cars;

run;

Here is the output.

Default Each distinct group value is represented in the plot by a different line color
and pattern. The line color is determined by the ContrastColor attribute of
the GraphData1–GraphDataN and GraphMissing style elements. The line
pattern is determined by the LineStyle attribute of the GraphData1–
GraphDataN and GraphMissing style elements.

Note The group values are mapped in the order in which they appear in the data.

Tip You can individually override the representations that are used to identify
the groups. For example, in some ODS styles, each distinct group value is
represented by a different line color and pattern. In that case, you can use
the PATTERN= setting on the LINEATTRS= option to assign the same line
pattern to all of the curves.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default TRUE
410 Chapter 6 • Plot Statements

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the MISSING= system option changes the default missing character, or
a user-defined format is applied to the group value. In those cases, the
attributes of the missing group value are determined by a GraphData1–
GraphDataN style element instead of by the GraphMissing style
element.

See “boolean ” on page 1339 for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the density curve.

Defaults For non-grouped data, the GraphFit style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements, and the
GraphFit:LineThickness style reference.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis.

Default VERTICAL

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE
DENSITYPLOT Statement 411

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Note The columns assigned to the X and Y roles are automatically

included in the data tip information.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Curve")

Default The column label or column name of the column assigned to the role.

Note The columns assigned to the X and Y roles are automatically included in
the data tip information.

WEIGHT=numeric-column | expression
specifies a column that contains a density-curve calculation a priori weight for each
observation of the input data object.

Requirement The value must be nonnegative.

Interaction If the value for an observation is missing or is less than 1, then the
observation is removed from the analysis.
412 Chapter 6 • Plot Statements

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Distribution Options
KERNEL (<C=positive-number> <WEIGHTFUNCTION=NORMAL |
QUADRATIC | TRIANGULAR> <MAXPOINTS=positive-integer>)
specifies a nonparametric kernel density estimate. The general form of the kernel
density estimator is as follows.
n x − xi
fλ x =
100h %
nλ ∑ K0 λ
i=1

In the equation, K 0 ⋅ is the weight function, λ is the bandwidth, n is the sample

size, and xi is the ith observation. You can use the C= suboption to specify the
bandwidth and the WEIGHT= suboption to specify the weight function K 0 ⋅ .

For more information, see the discussion of Kernel Density Estimates for the
UNIVARIATE procedure in the documentation for Base SAS.
TIP Use the KERNEL distribution option when the data density is not known.
C=positive-number
specifies a positive number (0 < number <= 100) that represents the standardized
bandwidth.
The value of λ, referred to as the bandwidth parameter, determines the degree of
smoothness in the estimated density function. You specify λ indirectly by
specifying a standardized bandwidth c with the C=kernel-option. If Q is the
interquartile range and n is the sample size, then c is related to λ by the following
equation:
1
λ = cQn − 5

Default Calculated from the data as the bandwidth that minimizes the
approximate mean integrated square error (MISE).

Range 0 to 100 (inclusive)

DENSITYPLOT Statement 413

WEIGHTFUNCTION=NORMAL | QUADRATIC | TRIANGULAR

specifies one of the weight functions NORMAL, QUADRATIC, or
TRIANGULAR.
The formulas for the weight functions are as follows:
NORMAL
1 1
K0 t = exp − 2 t2 for ‐∞ < t < ∞
2π

QUADRATIC
3
K0 t = 4
1 − t2 for t ≤ 1

TRIANGULAR
K 0 t = 1 − t for t ≤ 1

Default NORMAL

Note In prior SAS releases, the weight function was specified with the
WEIGHT= option. In SAS 9.4 and later releases, the WEIGHT= option
is not valid as a distribution option. You must use the
WEIGHTFUNCTION= option instead.

MAXPOINTS=positive-integer
specifies the maximum number of points generated for the curve.

Default 512

NORMAL (<MU=number> <SIGMA=number> <MAXPOINTS=number>)

specifies a normal density estimate, with mean and standard deviation. The fitted
density function equation is as follows:
100h % 1 x−μ 2
px = exp − 2 σ
for ‐∞ < x < ∞
σ 2π

In the equation, μ is the mean and σ is the standard deviation (σ > 0). You can
specify μ with the MU= suboption and σ with the SIGMA= suboption. By default,
ODS estimates μ with the sample mean and σ with the sample standard deviation.
For more information, see the discussion of Kernel Density Estimates for the
UNIVARIATE procedure in the documentation for Base SAS.
MU=number
specifies the mean.

Default The value is calculated from the data.

SIGMA=number
specifies the standard deviation.

Default The value is calculated from the data.

MAXPOINTS=number
specifies the maximum number of points generated for the curve.
Default: 200
414 Chapter 6 • Plot Statements

Details
A typical DENSITYPLOT statement specifies either the NORMAL or the KERNEL
distribution option. If no distribution option is specified, then the NORMAL() option is
used. The following syntax explicitly shows the default case:
DENSITYPLOT numeric-column / NORMAL()

To specify a kernel distribution, use the following plot syntax:

DENSITYPLOT numeric-column / KERNEL()

If more than one distribution option is specified, then the last distribution option
specified is used.

Examples

Example 1: DENSITYPLOT Statement

The following graph was generated by the “Example Program” on page 414:

Example Program
When used as a stand-alone plot or overlaid with other density plots, the dependent axis
shows the computed density values.
proc template;
define statgraph densityplot1;
begingraph;
entrytitle "Fitted Density Curves";
entrytitle "of Patient Weight";
entryfootnote halign=left "Framingham Heart Study";
layout overlay;
densityplot weight / normal()
lineattrs=graphfit name="n" legendlabel="Normal";
Example 2: Density Plot and Histogram 415

densityplot weight / kernel()

lineattrs=graphfit2 name="k" legendlabel="Kernel";
discretelegend "n" "k";
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.heart template=densityplot1;
label weight="Patient Weight";
run;

Example 2: Density Plot and Histogram

When one or more density plots are overlaid on a histogram, the dependent axis shows
the statistic indicated by the histogram’s SCALE= option. The area under each density
curve is equal to the area of the histogram. The following graph was generated by the
“Example Program” on page 415:

Example Program
proc template;
define statgraph densityplot2;
begingraph;
entrytitle "Patient Weight Distribution";
entrytitle "with Fitted Normal Curve";
entryfootnote halign=left "Framingham Heart Study";
layout overlay;
histogram weight / primary=true scale=count;
densityplot weight / normal() lineattrs=graphfit;
endlayout;
endgraph;
end;
run;
416 Chapter 6 • Plot Statements

proc sgrender data=sashelp.heart template=densityplot2;

label weight="Patient Weight";
run;

DROPLINE Statement
Creates a horizontal or vertical drop line from a point to an axis.
Requirement: A DROPLINE statement must be used within a 2-D layout (for example, an
OVERLAY, OVERLAYEQUATED, DATALATTICE, or DATAPANEL layout).

Syntax
DROPLINE X=x-axis-value | column | expression
Y=y-axis-value | column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the data for the line are considered when determining the
data ranges for the axes.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the drop line.
DATATRANSPARENCY=number
specifies the degree of the transparency of the drop line.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the drop line.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset all drop lines from discrete X values, or Y
values, or both.
DROPTO=X | Y | BOTH
specifies the axis to which the line is dropped.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
LABEL="string" | string-column
specifies a label for the point(s) indicated by the X= and Y= arguments.
LABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the drop line label(s).
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

DROPLINE Statement 417

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=x-axis-value | column | expression
specifies the X coordinate of the drop line(s).

Requirement Values must agree in type with the X-axis data type. For example,
you should use numeric SAS date or time values (or SAS date/time
constants) for a time axis.

Note When a character value is specified, leading blanks are honored and
trailing blanks are ignored when the specified value is compared with
the data values.

Tip By default, if a specified value is outside of the X-axis data range,

then the data range is extended to include the value. This behavior
can be changed with the CLIP= option.

Y=y-axis-value | column | expression

specifies the Y coordinate of the drop line(s).

Requirement Values must agree in type with the Y-axis data type.

Note When a character value is specified, leading blanks are honored and
trailing blanks are ignored when the specified value is compared with
the data values.

Tip By default, if a specified value is outside of the Y-axis data range,

then the data range is extended to include the value. This behavior
can be changed with the CLIP= option.

Optional Arguments
CLIP=TRUE | FALSE
specifies whether the data for the line are considered when determining the data
ranges for the axes.
FALSE
specifies that the reference line values are to be considered when the axis range is
determined. The reference lines are drawn as follows based on the axis type:
• For a discrete axis, the reference line values that are not already on the axis
are added to the end of the axis data list. When applicable, the axis values are
then sorted:
• If the axis values are numeric values, then they are sorted ordinally.
• If the axis values are character values and a sorting option is applied to the
axis, then they are sorted as specified by the sorting option.
Reference lines are then drawn at the specified locations.
• For a linear, log, or time axis, a new axis data list is created by performing a
mathematical union of the data values and the reference line values. The
reference lines are then drawn at the locations specified.
418 Chapter 6 • Plot Statements

TRUE
specifies that the reference line values are not to be considered when the axis
range is determined. The reference lines are drawn as follows based on the axis
type:
• For a discrete axis, if the reference line value exactly matches a value on the
axis, then a reference line is drawn at that location. Otherwise, the reference
line is not drawn.
Note: If the axis values are formatted, then the reference line value must
exactly match the formatted axis value in order for the line to be drawn.
• For a linear, log, or time axis, if the reference line value is within the axis
data range, then the reference line is drawn at the specified location.
Otherwise, the reference line is not drawn.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the drop line. The following figure shows drop
lines with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the GraphSkins:DataSkin
style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interaction This option overrides the BEGINGRAPH statement DATASKIN=

option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the drop line.
DROPLINE Statement 419

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the point labels.

DISCRETEOFFSET=number
specifies an amount to offset all drop lines from discrete X values, or Y values, or
both.

Default 0 (no offset, all drop lines are centered on discrete X values, or discrete
Y values, or both)

Range -0.5 to +0.5 where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If the
Y axis is discrete, then a positive offset is up. If REVERSE=TRUE on
the X or Y axis, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DROPTO=X | Y | BOTH
specifies the axis to which the line is dropped.
X
draws one or more droplines to an X axis.

Tip The XAXIS= option determines whether the X axis or X2 axis is the
endpoint for the line.

Y
draws one or more droplines to a Y axis.

Tip The YAXIS= option determines whether the Y axis or Y2 axis is the
endpoint for the line.

BOTH
draws one or more droplines to both axes.

Note This option is valid in the first maintenance release of SAS 9.4 and later
releases.

Default X

LABEL="string" | string-column
specifies a label for the point(s) indicated by the X= and Y= arguments.

Default No label is specified

Interaction If drawing multiple drop lines using X=column or Y=column, then you
can assign corresponding labels by using a column to define the labels.
420 Chapter 6 • Plot Statements

Note Starting with the first maintenance release of SAS 9.4, space is reserved
at the maximum end of the X axis to accommodate the length of the
labels regardless of where the labels appear in the plot.

Tips You can use the OFFSETMAX= axis option to adjust the amount of
space that is reserved on the X axis for the labels.

The font and color attributes for the label are specified by the
LABELATTRS= option.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the drop line label(s).

Default The GraphValueText style element.

Interaction For this option to have any effect, the LABEL= option must also be
specified.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the drop line.

Default The GraphReference style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X
Example: DROPLINE Statement 421

Restriction Another plot that establishes a data range for the designed axis must be
included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Restriction Another plot that establishes a data range for the designed axis must be
included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
A drop line is always drawn perpendicular from the specified point to the X (bottom),
X2 (top), Y (left), or Y2 (right) axis. Axis offsets do not apply to drop lines, so they
always meet the axis line.
The DROPTO= option controls whether a horizontal or vertical drop line is created.
DROPTO=X specifies the horizontal axis for a vertical drop line, and DROPTO=Y
specifies the vertical axis for a horizontal drop line.
A DROPLINE statement can be used only within a 2-D overlay-type layout (OVERLAY,
OVERLAYEQUATED, PROTOTYPE). Another plot statement that is derived from data
values that provide boundaries for the axis area must be included in the layout. For
example, it can be used with a scatter plot or a histogram. You can generate multiple
drop lines by specifying a column for X and Y. The column type (numeric or string)
must agree with the type of data presented on the axis.
To generate both a vertical and a horizontal drop line from a single point, use multiple
DROPLINE statements.

Example: DROPLINE Statement

Example Graph
The following graph was generated by the “Example Program” on page 422. The graph
shows two DROPLINE statements originating from the same point (X=3, Y=5). One
statement uses DROPTO=X and the other uses DROPTO=Y.
422 Chapter 6 • Plot Statements

Example Program
proc template;
define statgraph dropline;
begingraph;
entrytitle "Drop lines at Inflection Point";
layout overlay / yaxisopts=(linearopts=(viewmin=0));
seriesplot x=x y=y;
dropline x=3 y=5 / dropto=x
lineattrs=(color=blue pattern=dot) label="(3,5)";
dropline x=3 y=5 / dropto=y
lineattrs=(color=blue pattern=dot);
endlayout;
endgraph;
end;
run;

data test;
do X=0 to 8 by 0.25;
Y=(x-3)*(x-3) + 5;
output;
end;
run;

proc sgrender data=test template=dropline;

run;

ELLIPSE Statement
Creates a confidence ellipse computed from input data.
ELLIPSE Statement 423

Requirements: An ELLIPSE statement must be used in a two-dimensional overlay-type layout (for

example, an OVERLAY, OVERLAYEQUATED, or PROTOTYPE layout).
The ELLIPSE statement must be overlaid with another plot that is derived from data
values that provide boundaries for the axis area. It is typically overlaid with a scatter
plot.

Syntax
ELLIPSE X=numeric-column | expression
Y=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the data for the ellipse are considered when determining the
data ranges for the axes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the ellipse fill color and outline.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display an outlined ellipse, a filled ellipse, or an outlined
and filled ellipse.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the ellipse.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the ellipse outline.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate ellipse for each unique group value in the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
424 Chapter 6 • Plot Statements

ALPHA=positive-number
sets a significance value for the confidence level to compute for the ellipse.
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation
that is read.
TYPE=MEAN | PREDICTED
specifies the type of ellipse.

Required Arguments
X=numeric-column | expression
specifies the numeric column for the X values.
Y=numeric-column | expression
specifies the numeric column for the Y values.

Optional Arguments
ALPHA=positive-number
sets a significance value for the confidence level to compute for the ellipse.

Default 0.05

Range 0 < number < 1

Note ALPHA=0.05 represents a 95% confidence level.

See TYPE=

CLIP=TRUE | FALSE
specifies whether the data for the ellipse are considered when determining the data
ranges for the axes.
FALSE
The data for the ellipse contribute to the data range for each axis. Each axis
might be extended to force the display of the entire ellipse.
TRUE
The data for the ellipse are ignored when establishing axis scales. Each axis scale
is determined by the other plots in the parent layout. This might result in the
ellipse not being entirely displayed (clipped) if its data range is not within the
data ranges of the other plots.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

DATATRANSPARENCY=number
specifies the degree of the transparency of the ellipse fill color and outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
ellipse fill. You can combine this option with FILLATTRS= to set one
transparency for the ellipse outline but a different transparency for the
ellipse fill. Example:
ELLIPSE Statement 425

datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display an outlined ellipse, a filled ellipse, or an outlined and
filled ellipse.
STANDARD
displays an outlined, unfilled ellipse
ALL
displays an outlined, filled ellipse
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays an outlined ellipse

FILL displays a filled ellipse

Default GraphEllipse:DisplayOpts style reference.

Tip Use FILLATTRS= and OUTLINEATTRS= to control the appearance of the

ellipse.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the ellipse.

Default The GraphDataDefault style element.

Interaction For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

Tip The DATATRANSPARENCY= option sets the transparency for the

ellipse fill and ellipse outline. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outline but a
different transparency for the fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that is
read.

Default All observations have a frequency count of 1.

Restriction If the value of the numeric-column is missing or is less than 1, then the
observation is not used in the analysis. If the value is not an integer,
then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation.
426 Chapter 6 • Plot Statements

GROUP=column | discrete-attr-var | expression

creates a separate ellipse for each unique group value in the specified column.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

For example, the Sashelp.Iris data contains a column named Species, which
identifies the species of each plant. This column could be used in the ELLIPSE
statement to group the ellipses in the display:
proc template;
define statgraph ellipsegroup;
begingraph;
entrytitle "95% Prediction By Species";
layout overlay;
scatterplot x=petallength y=petalwidth / name="sp"
group=species;
ellipse x=petallength y=petalwidth / group=species
type=predicted alpha=0.05 name="p95";
mergedlegend "sp" "p95" / location=inside
autoalign=(topleft) across=1;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.iris template=ellipsegroup;

run;

Here is the output.

Default Each distinct group value is represented in the plot by a different line color
and pattern. The line color is determined by the ContrastColor attribute of
ELLIPSE Statement 427

the GraphData1–GraphDataN and GraphMissing style elements. The line

pattern is determined by the LineStyle attribute of the GraphData1–
GraphDataN and GraphMissing style elements.

Note The group values are mapped in the order in which they appear in the data.

Tip You can individually override the representations that are used to identify
the groups. For example, in some ODS styles, each distinct group value is
represented by a different line color and pattern. In that case, you can use
the PATTERN= setting on the OUTLINEATTRS= option to assign the
same line pattern to all of the ellipses.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the MISSING= system option changes the default missing character, or
a user-defined format is applied to the group value. In those cases, the
attributes of the missing group value are determined by a GraphData1–
GraphDataN style element instead of by the GraphMissing style
element.

See “boolean ” on page 1339 for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the ellipse outline.

Defaults For non-grouped data, the GraphDataDefault style element.

428 Chapter 6 • Plot Statements

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements, and the LineThickness
attribute of the GraphDataDefault style element.

Interaction For this option to have any effect, the outlines must be enabled by the
ODS style or the DISPLAY= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

TYPE=MEAN | PREDICTED
specifies the type of ellipse.
MEAN
specifies a confidence ellipse of the mean
PREDICTED
specifies a prediction ellipse for a new observation

Default MEAN

See ALPHA= option for specifying a confidence level.

For statistical details about how the ellipse is calculated, see “Confidence
and Prediction Ellipses ” on page 429.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details

Statement Description
The ELLIPSE statement can be used only within 2-D overlay-type layouts. It computes
an ellipse for a set of points specified by the X and Y columns and a confidence level
specified by the ALPHA= option. Use the TYPE= option to control whether a predicted
or confidence ellipse is generated.
Example: ELLIPSE Statement 429

Confidence and Prediction Ellipses

Two types of ellipses can be computed for the input data (where observations correspond
to points in a scatter plot). One is a confidence ellipse for the population mean
(TYPE=MEAN), and the other is a prediction ellipse for a new observation
(TYPE=PREDICT). Both assume a bivariate normal distribution.
Let Z and S be the sample mean and sample covariance matrix of a random sample of
size n from a bivariate normal distribution with mean μ and covariance matrix Σ. The
variable Z − μ is distributed as a bivariate normal variate with mean zero and covariance
1
( n )Σ, and it is independent of S. Using Hotelling’s T 2 statistic, which is defined as
follows:

T 2 = n(Z − μ)′S−1(Z − μ)

A 100(1 − α) % confidence ellipse for μ is computed from the following equation:

n 2
n−1
(Z − μ)′S−1(Z − μ) = F
n − 2 2, n − 2
(1 − α)

where F2, n − 2(1 − α) is the (1 − α) critical value of an F distribution with degrees of

freedom 2 and n − 2.
A prediction ellipse is a region for predicting a new observation in the population. It also
approximates a region containing a specified percentage of the population.
Denote a new observation as the bivariate random variable Z new. The following variable:

Z new − Z = (Z new − μ) − (Z − μ)

is distributed as a bivariate normal variate with mean zero (the zero vector) and
1
covariance (1 + n )Σ, and it is independent of S. A 100(1 − α) % prediction ellipse is then
given by the following equation:
n 2(n + 1)
n−1
(Z − μ)′S−1(Z − μ) = n−2
F2, n − 2(1 − α)

The family of ellipses generated by different critical values of the F distribution has a
common center (the sample mean) and common major and minor axis directions.
The shape of an ellipse depends on the aspect ratio of the plot. The ellipse indicates the
correlation between the two variables if the variables are standardized (by dividing the
variables by their respective standard deviations). In this situation, the ratio between the
major and minor axis lengths is:
1+ r
1− r

In particular, if r = 0, then the ratio is 1, which corresponds to a circular confidence

contour and indicates that the variables are uncorrelated. A larger value of the ratio
indicates a larger positive or negative correlation between the variables.

Example: ELLIPSE Statement

The following graph was generated by the “Example Program” on page 430:
430 Chapter 6 • Plot Statements

Example Program
This example overlays two ELLIPSE statements on a SCATTERPLOT of the same data.
Both ELLIPSE statements use TYPE= PREDICTED. One ELLIPSE statement uses
ALPHA= 0.2, and the other uses ALPHA=0.05.
proc template;
define statgraph ellipse;
begingraph;
entrytitle "Prediction Ellipses";
layout overlayequated / equatetype=equate;
scatterplot x=petallength y=petalwidth /
datatransparency=0.5;
ellipse x=petallength y=petalwidth /
type=predicted alpha=0.2
name="p80" legendlabel="80%"
outlineattrs=graphconfidence;
ellipse x=petallength y=petalwidth /
type=predicted alpha=0.05
name="p95" legendlabel="95%"
outlineattrs=graphconfidence2;
discretelegend "p80" "p95" /
location=inside autoalign=(topleft);
endlayout;
entryfootnote halign=left "Fisher's Iris Data";
endgraph;
end;
run;
proc sgrender data=sashelp.iris template=ellipse;
run;
ELLIPSEPARM Statement 431

ELLIPSEPARM Statement
Creates an ellipse specified by slope, axis, and origin parameters.
Requirements: An ELLIPSEPARM statement must be used in a two-dimensional overlay-type layout
(for example, an OVERLAY, OVERLAYEQUATED, or PROTOTYPE layout).
The ELLIPSEPARM statement must be overlaid with another plot that is derived from
data values that provide boundaries for the axis area. It is typically overlaid with a
scatter plot.
Tip: You can generate a single ellipse by specifying a constant for each required
argument. You can generate multiple ellipses by specifying a numeric column for any
or all required arguments. If any of the SEMIMAJOR= , SEMIMINOR= , XORIGIN=
or YORIGIN= constants or columns contains a missing value, then no ellipse is
drawn. To request a vertical major axis, specify SLOPE=. (missing value) as a
constant or column value.

Syntax
ELLIPSEPARM SEMIMAJOR=number | numeric-column | expression
SEMIMINOR=number | numeric-column | expression
SLOPE=number | numeric-column | expression
XORIGIN=number | numeric-column | expression
YORIGIN=number | numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the data for the ellipse are considered when determining the
data ranges for the axes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the ellipse fill color and outline.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display an outlined ellipse, a filled ellipse, or an outlined
and filled ellipse.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the ellipse.
INDEX=positive-integer-column | expression
specifies indices for mapping ellipse attributes (fill and outline) to one of the
GraphData1–GraphDataN style elements.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the ellipse outline.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
432 Chapter 6 • Plot Statements

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate ellipse for each unique group value of the specified
column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
SEMIMAJOR=number | numeric-column | expression
specifies half the length of the major axis for the ellipse. Values are in the same units
as the data. The SEMIMAJOR value can be greater than, smaller than, or equal to
the SEMIMINOR value.

Restriction If a numeric column is specified and the column contains negative

values, then no ellipses are drawn. In the second maintenance release of
SAS 9.4 and in earlier releases, the same is true if the column contains
missing values. Starting with the third maintenance release of SAS 9.4,
if the column contains a missing value, the ellipse for that observation
is not drawn. However, the remaining ellipses that have nonmissing
values for this argument and for SEMIMINOR= are drawn.

SEMIMINOR=number | numeric-column | expression

specifies half the length of the minor axis for the ellipse. Values are in the same units
as the data. The SEMIMINOR value can be greater than, smaller than, or equal to the
SEMIMAJOR value.

Restriction If a numeric column is specified and the column contains negative

values, then no ellipses are drawn. In the second maintenance release of
SAS 9.4 and in earlier releases, the same is true if the column contains
missing values. Starting with the third maintenance release of SAS 9.4,
if the column contains a missing value, the ellipse for that observation
is not drawn. However, the remaining ellipses that have nonmissing
values for this argument and for SEMIMAJOR= are drawn.

SLOPE=number | numeric-column | expression

specifies the slope of the major axis for the ellipse. Slope can be positive or negative.

Note The slope value is in the data space and might or might not be maintained in
the screen space. Thus, setting SLOPE=1 does not always generate a 45
degree line on the screen.

Tip Setting SLOPE=0 creates a major axis parallel to the X-axis. Setting
SLOPE=. (missing value) creates a major axis parallel to the Y-axis.
ELLIPSEPARM Statement 433

XORIGIN=number | numeric-column | expression

specifies the X coordinate of the center of the ellipse. Values are in the units of the
data.

Tip By default, if the value specified for the XORIGIN= option is outside of the X-
axis data range, then the data range is extended to include the specified point.
This behavior can be changed with the CLIP= option.

YORIGIN=number | numeric-column | expression

specifies the Y coordinate of the center of the ellipse. Values are in the units of the
data.

Tip By default, if the value specified for the YORIGIN= option is outside of the Y-
axis data range, then the data range is extended to include the specified point.
This behavior can be changed with the CLIP= option.

Optional Arguments
CLIP=TRUE | FALSE
specifies whether the data for the ellipse are considered when determining the data
ranges for the axes.
FALSE
The data for the ellipse contribute to the data range for each axis. Each axis
might be extended to force the display of the entire ellipse.
TRUE
The data for the ellipse are ignored when establishing axis scales. Each axis scale
is determined by the other plots in the parent layout. This might result in the
ellipse not being entirely displayed (clipped) if its data range is not within the
data ranges of the other plots.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

DATATRANSPARENCY=number
specifies the degree of the transparency of the ellipse fill color and outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
ellipse fill. You can combine this option with FILLATTRS= to set one
transparency for the ellipse outline but a different transparency for the
ellipse fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display an outlined ellipse, a filled ellipse, or an outlined and
filled ellipse.
STANDARD
displays an outlined, unfilled ellipse
ALL
displays an outlined, filled ellipse
434 Chapter 6 • Plot Statements

(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays an outlined ellipse

FILL displays a filled ellipse

Default GraphEllipse:DisplayOpts style reference.

Tip Use FILLATTRS= and OUTLINEATTRS= to control the appearance of the

ellipse.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the ellipse.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the GraphData1–GraphDataN style elements.

Interaction For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

Tip The DATATRANSPARENCY option sets the transparency for the

ellipse fill and ellipse outline. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outline but a
different transparency for the fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

GROUP=column | discrete-attr-var | expression

creates a separate ellipse for each unique group value of the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

The representations that are used to identify the groups can be overridden
individually. For example, each distinct group value might be represented by a
different line pattern for the ellipses, but the PATTERN= suboption of the
OUTLINEATTRS= option could be used to assign the same line pattern to all ellipse
outlines, letting outline color distinguish group values.

Defaults If DISPLAY= (OUTLINE), then each distinct group value might be

represented in the plot by a different combination of outline color and
line pattern. Line color and pattern vary according to the ContrastColor
and LineStyle attributes of the GraphData1–GraphDataN and
GraphMissing style elements. Line thickness (for grouped and
ungrouped data) is controlled by the OUTLINEATTRS= option.
ELLIPSEPARM Statement 435

If DISPLAY=(FILL), then each distinct group value might be

represented in the plot by a different fill color defined by the Color
attribute of the GraphData1–GraphDataN and GraphMissing style
elements.

If DISPLAY=(FILL OUTLINE), then each distinct group value might

be represented in the plot by a different fill color, outline color, and
outline pattern.

Restriction When the GROUP= option is specified, the group value must be a
character or numeric column. For each group value, there must be a
numeric column that does not contain missing values for
SEMIMAJOR=, SEMIMINOR=, XORIGIN=, and YORIGIN=.The
SLOPE=column can contain missing values. Under these
circumstances, an ellipse is drawn for each group value.

Interaction The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping ellipse attributes (fill and outline) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.
436 Chapter 6 • Plot Statements

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the ellipse outline.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Restriction Another plot that establishes a data range for the designed axis must be
included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.
Example: ELLIPSEPARM Statement 437

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Restriction Another plot that establishes a data range for the designed axis must be
included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
The ELLIPSEPARM statement plots an ellipse with specified semimajor and semiminor
axis lengths, a specified slope for the major axis, and the ellipse center. The ellipse
center is the point of intersection of the semimajor and semiminor axes. It can be used
only within a two-dimensional overlay-type layout (for example, an OVERLAY,
OVERLAYEQUATED, or PROTOTYPE layout). The ELLIPSEPARM statement does
not perform computations on input data to derive the location and shape of the ellipse.
Another plot that is derived from data values that provide boundaries for the axis area
must be included in the layout. It is typically overlaid with a scatter plot. See “ELLIPSE
Statement” on page 422 for information about computed ellipses.

Example: ELLIPSEPARM Statement

The following graph was generated by the “Example Program” on page 438:
438 Chapter 6 • Plot Statements

Overview
This is a simplified version of the CorrLoadPlot template for PROC PLS in the
SAS/STAT product. It consists of overlaid scatter plots of the scores of the first two
factors, the loadings of the model effects, and the loadings of the dependent variables.
The loadings are scaled so that the amount of variation in the variables that is explained
by the model is proportional to the distance from the origin; circles indicating various
levels of explained variation are also overlaid.
The circles are drawn with ELLIPSEPARM statements by setting the SEMIMAJOR and
SEMIMINOR lengths to be the same. Notice that all circles are concentric because they
share the same origin (0,0). The radius of each circle is passed dynamically.
Also note that an OVERLAYEQUATED layout was used to force the length of unit
intervals on both axes to be the same.
The input data shown is representative of that computed by PROC PLS for the
Correlation Loadings Plot. For more details, see the first example for PROC PLS in the
SAS/STAT user’s guide.

Example Program
Here is the SAS program for this example.
proc template;
define statgraph ellipseparm;
dynamic RADIUS1 RADIUS2 RADIUS3 RADIUS4 ;
begingraph;
entrytitle "Correlation Loading Plot";
layout overlayequated / equatetype=square
commonaxisopts=
(tickvaluelist=(-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1)
viewmin=-1 viewmax=1)
xaxisopts=
(label="Factor 1" offsetmin=0.05 offsetmax=0.05)
yaxisopts=
(label="Factor 2" offsetmin=0.05 offsetmax=0.05);
ellipseparm semimajor=RADIUS1 semiminor=RADIUS1 slope=0
xorigin=0 yorigin=0 / clip=true display=(outline)
outlineattrs=(pattern=dash) datatransparency=0.75;
ellipseparm semimajor=RADIUS2 semiminor=RADIUS2 slope=0
xorigin=0 yorigin=0 / clip=true display=(outline)
outlineattrs=(pattern=dash) datatransparency=0.75;
ellipseparm semimajor=RADIUS3 semiminor=RADIUS3 slope=0
xorigin=0 yorigin=0 / clip=true display=(outline)
outlineattrs=(pattern=dash) datatransparency=0.75;
ellipseparm semimajor=RADIUS4 semiminor=RADIUS4 slope=0
xorigin=0 yorigin=0 / clip=true display=(outline)
outlineattrs=(pattern=dash) datatransparency=0.75;

scatterplot x=xcirclelabel y=ycirclelabel / primary=true

markercharacter=circlelabel datatransparency=0.75 ;
scatterplot x=corr1 y=corr2 / name="ScatterVars"
group=corrgroup markercharacter=corrlabel;
discretelegend "ScatterVars";
endlayout;
endGraph;
end;
FRINGEPLOT Statement 439

run;

data corrplot;
infile cards missover;
input Corr1 Corr2 CorrGroup &$18. CorrLabel :$7.
xCircleLabel yCircleLabel CircleLabel :$8.;
datalines;
-0.179 -0.268 Predictor Loading S1 0 0.5 25%
0.105 0.332 Predictor Loading S2 0 -0.5 25%
-0.654 0.094 Predictor Loading S3 0 0.707 50%
-0.653 0.685 Predictor Loading S4 0 -0.707 50%
0.096 0.059 Predictor Loading S5 0 0.866 75%
0.132 0.036 Predictor Loading L1 0 -0.866 75%
0.087 0.156 Predictor Loading L2 0 1 100%
0.940 0.160 Predictor Loading L3 0 -1 100%
0.607 -0.350 Predictor Loading L4
0.096 0.059 Predictor Loading L5
-0.111 -0.534 Predictor Loading P1
0.003 0.256 Predictor Loading P2
0.293 0.551 Predictor Loading P3
-0.480 0.643 Predictor Loading P4
-0.096 -0.059 Predictor Loading P5
0.946 0.279 Response Loading log_RAI
-0.196 0.403 Observation 1
0.020 -0.001 Observation 2
-0.195 0.324 Observation 3
0.021 -0.079 Observation 4
-0.009 -0.274 Observation 5
0.567 0.294 Observation 6
-0.096 -0.059 Observation 7
0.258 0.210 Observation 8
-0.104 -0.309 Observation 9
-0.187 -0.458 Observation 10
0.051 -0.078 Observation 11
0.017 -0.260 Observation 12
-0.621 0.372 Observation 13
0.392 0.138 Observation 14
0.080 -0.221 Observation 15
run;

proc sgrender data=corrplot template=ellipseparm;

dynamic radius1=0.50 radius2=0.71 radius3=0.87 radius4=1;
run;

FRINGEPLOT Statement
Creates a fringe plot on the X axis of an X-Y plot.

Syntax
FRINGEPLOT numeric-column | expression </option(s)>;
440 Chapter 6 • Plot Statements

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the fringe.
FRINGEHEIGHT=dimension
specifies the height of the fringe lines.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the fringe lines for the data points.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
fringe line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of lines for each unique group value of the specified
column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Argument
numeric-column | expression
specifies a column that provides the X coordinates of the data values.
FRINGEPLOT Statement 441

Optional Arguments
DATATRANSPARENCY=number
specifies the degree of the transparency of the fringe.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

FRINGEHEIGHT=dimension
specifies the height of the fringe lines.

Default 10 px

See “dimension” on page 1340

GROUP=column | discrete-attr-var | expression

creates a distinct set of lines for each unique group value of the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value is represented in the plot by a different

color. The colors are determined by the ContrastColor attribute of the
GraphData1–GraphDataN and GraphMissing style elements. The line
style and line thickness are determined by the GraphDataDefault style
element.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of colors and line
patterns.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Note You can override the representations that are used to identify the
groups. For example, you can use the
LINEATTRS=(PATTERN=pattern) option to assign the same line
pattern to all of the plot’s line patterns, letting line color indicate group
values. Likewise, you can use LINEATTRS=(COLOR=color) to
assign the same color to all lines, letting line pattern indicate group
values.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.
442 Chapter 6 • Plot Statements

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The X-column label. If a label is not defined, then the X-column name
is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the fringe lines for the data points.

Default The GraphDataDefault style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

FRINGEPLOT Statement 443

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined role X.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a fringe line.
If this option is used, then it replaces all of the information that is displayed by
default. Roles for columns that do not contribute to the fringe plot can be specified
along with roles that do.
role-list
an ordered, space-separated list of unique FRINGEPLOT and user-defined roles.
The FRINGEPLOT role is X.

Tip User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns assigned to
the roles X and Y, as well as the column Pct, which is not assigned to
any pre-defined FRINGEPLOT role. The Pct column should appear
first in the data tip:

ROLENAME=(TIP1=PCT)
TIP=(X TIP1)

NONE
suppresses data tips from the plot.

Default The column assigned to the X role is automatically included in the

data tip information.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
444 Chapter 6 • Plot Statements

specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.
Example: FRINGEPLOT Statement 445

Note The display of the fringe lines is always anchored on the X-axis (at the
bottom of the plot area), even when the FRINGEPLOT’s X-column
data range is mapped to the X2 axis.

Details
In a FRINGEPLOT, each fringe line represents the location of the corresponding raw
data value on the X axis. All fringe lines are of equal length.

Example: FRINGEPLOT Statement

The following graph was generated by the “Example Program” on page 445:

Example Program
proc template;
define statgraph fringeplot;
dynamic VAR VARLABEL;
begingraph;
entrytitle "Histogram and Fringeplot";
layout overlay / xaxisopts=(label=VARLABEL)
yaxisopts=(offsetmin=0.03);
fringeplot VAR / datatransparency=0.75
fringeheight=3pct;
histogram VAR;
endlayout;
endgraph;
end;
run;
446 Chapter 6 • Plot Statements

proc sgrender data=sashelp.cars template=fringeplot;

dynamic var="weight" varlabel="Weight (LBS)";
run;

HEATMAP Statement
Creates a plot of color-coded rectangles for the response variable of a pair of X and Y variables after it bins
the data in two dimensions.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Syntax
HEATMAP X=numeric-column | expression
Y=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies a numeric column that is used to color the heat map regions.
COLORSTAT= FREQ | PCT | PROPORTION | SUM | MEAN
specifies the statistic to be calculated for the COLORRESPONSE= column.
DATATRANSPARENCY=number
specifies the degree of the transparency of the outline and fill for each region.
DISPLAY=ALL | STANDARD | (display-options)
specifies the degree of the transparency of the rectangles.
FILLATTRS=(TRANSPARENCY=number)
specifies the transparency of the interior fill area of the regions.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the rectangle outlines.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient that is specified by the ODS style
that is in effect or by the COLORMODEL= option.
WEIGHT=numeric-column
specifies a variable in the input data set that contains values to be used as a
priori weights for the colored-region calculations.
XGAP=number
specifies the amount of horizontal space on either side of each color-coded
region in the heat map.
YGAP=number
specifies the amount of horizontal space on either side of each color-coded
region in the heat map.

Axes options
PRIMARY=TRUE | FALSE
HEATMAP Statement 447

specifies that the data columns for this plot are used for determining default
axis features.
XAXIS=X | X2
specifies whether data is mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
XBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for x-axis tick marks.
XENDLABELS=TRUE | FALSE
specifies whether the X axis ticks and tick values are placed at the bin end-
points or at the bin mid-points.
YAXIS=Y | Y2
specifies whether data is mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
YBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for x-axis tick marks.
YENDLABELS=TRUE | FALSE
specifies whether the Y axis ticks and tick values are placed at the bin end-
points or at the bin mid-points.

Binning options
NXBINS=positive-integer
specifies the number of bins to use for the X role.
NYBINS=positive-integer
specifies the number of bins to use for the Y role.
XBINSIZE=positive-number
specifies the size of bins along the X role, in data units.
XBINSTART=number
specifies the data value for the first bin of the X role.
XBOUNDARY=UPPER | LOWER
specifies how an input value is counted when it lies on the endpoint of an X
bin.
XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the XBINSTART= value represents the midpoint, lower
endpoint, or upper endpoint of the bin.
YBINSIZE=positive-number
specifies the size of bins along the Y role, in data units.
YBINSTART=number
specifies the data value for the first bin of the Y role.
YBOUNDARY=UPPER | LOWER
specifies how an input value is counted when it lies on the endpoint of an Y
bin.
YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the YBINSTART= value represents the midpoint, lower
endpoint, or upper endpoint of the bin.

Data option
DISCRETEX=TRUE | FALSE
specifies whether the X axis is discrete when X= specifies a numeric column.
DISCRETEY=TRUE | FALSE
specifies whether the Y axis is discrete when Y= specifies a numeric column.
448 Chapter 6 • Plot Statements

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
rectangle.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Plot reference options

NAME="string"
assigns a name to a HEATMAP statement for reference in other template
statements and log messages.

Statistics options
FREQ=numeric-column | expression
specifies a column for the frequency count for each observation in the input
data.

Required Arguments
X=column | expression
specifies the column for the X values.

Requirement If X= specifies a numeric column, then the DISCRETEX= option

must be set correctly for the X-axis type. When it is not set correctly,
the heat map is not drawn. If the X-axis type is discrete, then you
must specify DISCRETEX=TRUE in the HEATMAP statement.
Otherwise, DISCRETEX=FALSE must be in effect.

See DISCRETEX= on page 450

Y=column | expression
specifies the column for the Y values.

Requirement If Y= specifies a numeric column, then the DISCRETEY= option

must be set correctly for the Y-axis type. When it is not set correctly,
the heat map is not drawn. If the Y-axis type is discrete, then you
must specify DISCRETEY=TRUE in the HEATMAP statement.
Otherwise, DISCRETEY=FALSE must be in effect.

See DISCRETEY= on page 450

Optional Arguments
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:
HEATMAP Statement 449

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to the
color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies a numeric column that is used to color the heat map regions.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Tips To display a legend with this option in effect, use a CONTINUOUSLEGEND

statement.

Use the COLORSTAT= option to specify the statistic to compute for the
COLORRESPONSE= column.

COLORSTAT= FREQ | PCT | PROPORTION | SUM | MEAN

specifies the statistic to be calculated for the COLORRESPONSE= column.

Default FREQ

Interaction When FREQ, PCT, or PROPORTION is specified, the colors are

controlled by the FREQ= and WEIGHT= option values. When SUM or
MEAN is specified, the colors are controlled by the FREQ=,
WEIGHT=, and COLORRESPONSE= option values.
450 Chapter 6 • Plot Statements

See COLORRESPONSE=

FREQ=

WEIGHT=

DATATRANSPARENCY=number
specifies the degree of the transparency of the outline and fill for each region.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip Use the FILLATTRS= option to set transparency for just the rectangle fills.
You can combine this option with FILLATTRS= to set one transparency for
the rectangle outlines and a different transparency for the rectangle fills.
Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEX=TRUE | FALSE
specifies whether the X axis is discrete when X= specifies a numeric column.

Default FALSE

Requirement If X= specifies a numeric column and the X axis type is discrete, then
you must specify DISCRETEX=TRUE. Otherwise, the heat map
might not be drawn.

Interaction If X= specifies a character column, then this option is ignored, and

the X axis is considered to be discrete.

See X= on page 448

DISCRETEY=TRUE | FALSE
specifies whether the Y axis is discrete when Y= specifies a numeric column.

Default FALSE

Requirement If Y= specifies a numeric column and the Y axis type is discrete, then
you must specify DISCRETEY=TRUE. Otherwise, the heat map
might not be drawn.

Interaction If Y= specifies a character column, then this option is ignored, and

the Y axis is considered to be discrete.

See Y= on page 448

DISPLAY=ALL | STANDARD | (display-options)

specifies the degree of the transparency of the rectangles.

STANDARD displays colored regions.

ALL displays outlined, colored regions.
display-options a space-separated list of options, enclosed in parentheses.
Currently, only OUTLINE is supported, which displays
outlined, filled rectangles.
HEATMAP Statement 451

Default STANDARD

FILLATTRS=(TRANSPARENCY=number)
specifies the transparency of the interior fill area of the regions.

Default The DATATRANSPARENCY= option value

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction Only the TRANSPARENCY= sub option is honored. If a style element

or any other fill sub option is specified, it is ignored by the HEATMAP
statement.

Interaction This option overrides the DATATRANSPARENCY= option only for

the colored region. It does not affect the region outlines.

Tip Use this option when you want the region fill to have a different degree
of transparency than the outline.

FREQ=numeric-column | expression
specifies a column for the frequency count for each observation in the input data.

Restrictions If the value of FREQ for a given observation is missing or is less than
1, that observation is not used in the analysis.

If the value is not an integer, only the integer portion is used.

NAME="string"
assigns a name to a HEATMAP statement for reference in other template statements
and log messages.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Tip This option is used mostly with legend statements in order to

coordinate the use of colors and line patterns between the graph and the
legend.

NXBINS=positive-integer
specifies the number of bins to use for the X role. The system determines the
XBINSIZE= and XBINSTART= values if they are not specified. The bins always
span the range of the data.

Default Determined by the system.

Requirement You must specify a value of 2 or greater. Otherwise, this option is

ignored.

NYBINS=positive-integer
specifies the number of bins to use for the Y role. The system determines the
YBINSIZE= and YBINSTART= values if they are not specified. The bins always
span the range of the data.

Default Determined by the system.

Requirement You must specify a value of 2 or greater. Otherwise, this option is

ignored.
452 Chapter 6 • Plot Statements

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the rectangle outlines.

Defaults For filled regions, the GraphOutlines style element

For unfilled regions, the GraphOutlinesUnfilled style element

Interaction For this option to have any effect, outlines must be enabled by the ODS
style or by the DISPLAY= option.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element value

“Line Options” on page 1349 for available line-options values

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot are used for determining default axis
features.

Default FALSE. If no plot in an OVERLAY layout is designated as primary,

the data columns associated with the first plot that could be primary is
considered primary on a per-axis basis.

Restrictions Only one plot in an overlay can be primary on a per-axis basis. If

multiple plots specify PRIMARY=TRUE for the same axis, the last
one specified is considered the primary plot.

This option is ignored if the plot is placed in a GRIDDED or

LATTICE layout block.

Tip This option is needed only when there are two or more plots within an
overlay that contribute to a common axis. If PRIMARY=TRUE for
one of them, then that plot’s data columns are used to determine the
axis features, regardless of where this plot statement occurs within the
OVERLAY layout block.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient that is specified by the ODS style that is in
effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL= on page 448

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
HEATMAP Statement 453

ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, COLORGROUP=, and
COLORRESPONSE=.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a rectangle. If
this option is used, the specified information replaces all the information that is
displayed by default. You can specify roles for columns that do not contribute to the
bar chart along with roles that do.
(role-list)
an ordered, space-separated list of unique HEATMAP roles and user-defined
roles. HEATMAP roles include X, Y, and the implicit count.

Example To display data tips for the columns assigned to the roles X and Y as
well as the user-defined role TIP1:
ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

NONE
suppresses data tips and URLs (if requested) in the graph output.

Default The columns assigned to these roles are automatically included in the
data tip information: X and Y

Requirement To generate data tips, you must include an ODS GRAPHICS ON

statement that specifies the IMAGEMAP option, and write the graphs
to the ODS HTML destination.

Tip You can control the labels and formats for the TIP variables with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
(role-format-list)
a space-separated list of role-name = format pairs.

Example To specify a format for the user-defined TIP1 role:

ROLENAME=(TIP1=PCT)
TIP=(TIP1 X Y)
TIPFORMAT=(TIP1=PERCENT7.2)

Default The column format of the variable assigned to the role or BEST6. if
no format is assigned to a numeric column.

Requirement This option provides a way to control the formats of columns that
appear in data tips. Only the roles that appear in the TIP= option are
used.

TIPLABEL=(role-label-list)
specifies display labels for tip columns.
454 Chapter 6 • Plot Statements

(role-label-list)
a space-separated list of role-name ="string" pairs.

Example To specify a label for the user-defined TIP1 role:

ROLENAME=(TIP1=PCT)
TIP=(TIP1 X Y)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the variable assigned to the
role.

Requirement This option provides a way to control the labels of columns that
appear in data tips. Only the roles that appear in the TIP= option are
used.

WEIGHT=numeric-column
specifies a variable in the input data set that contains values to be used as a priori
weights for the colored-region calculations.

Requirement The values of the weight variable must be nonnegative.

Note If an observation's weight is zero, negative, or missing, the

observation is deleted from the analysis.

XAXIS=X | X2
specifies whether data is mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

XBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for x-axis tick marks. When this option is
FALSE, a standard axis is used, ignoring bin boundaries and midpoints.

Default TRUE

Interactions When this option is TRUE, the XENDLABELS= option determines

how the axis ticks and value labels are displayed. When this option is
FALSE, the XENDLABELS= option is ignored.

When this option is TRUE, the system attempts to compute nice

values for the bin end points. In that case, the sum of the bin data
ranges might exceed the actual data range.

When this option is TRUE, the axis ticks are in predetermined

locations and are not changed when the parent layout specifies axis
suboptions such as TICKVALUELIST=, TICKVALUESEQUENCE=,
and INCLUDERANGES=.

See “boolean ” on page 1339 for other Boolean values that you can use.
HEATMAP Statement 455

XBINSIZE=positive-number
specifies the size of bins along the X role, in data units. The system determines the
NXBINS= and XBINSTART= values if they are not specified. The bins always span
the X data range.

Default Determined by the system.

Interaction The XGAP= option is applied after this option.

XBINSTART=number
specifies the data value for the first bin of the X role. The system determines the
NXBINS= and XBINSIZE= values if they are not specified. The bins always span
the X data range.

Default Determined by the system.

Interaction The XVALUES= option specifies how this value is interpreted.

XBOUNDARY=UPPER | LOWER
specifies how an input value is counted when it lies on the endpoint of an X bin. If
this option is set to UPPER, then the value is counted as one of the values in the
upper bin. Otherwise, it is counted in the lower bin.

Default UPPER

XENDLABELS=TRUE | FALSE
specifies whether the X axis ticks and tick values are placed at the bin end-points or
at the bin mid-points.

Default FALSE. The axis ticks and tick values are placed at the bin mid-
points.

Interactions If XBINAXIS=FALSE, then this option is ignored.

This option is ignored when the X column is non-numeric.

Tip The axis ticks and tick value placements are independent of the
XVALUES= option.

XGAP=number
specifies the amount of horizontal space on either side of each color-coded region in
the heat map.

Default 0

Interaction If the XBINSIZE= option is specified, the XGAP= value is subtracted

from the specified X bin size.

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the XBINSTART= value represents the midpoint, lower endpoint,
or upper endpoint of the bin.

Default MIDPOINTS

See XBINSTART= on page 455

XENDLABELS= on page 455

456 Chapter 6 • Plot Statements

YAXIS=Y | Y2
specifies whether data is mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for x-axis tick marks. When this option is
FALSE, a standard axis is used, ignoring bin boundaries and midpoints.

Default TRUE

Interactions When this option is TRUE, the YENDLABELS= option determines

how the axis ticks and value labels are displayed. When this option is
FALSE, the YENDLABELS= option is ignored.

When this option is TRUE, the system attempts to compute nice

values for the bin end points. In that case, the sum of the bin data
ranges might exceed the actual data range.

When this option is TRUE, the axis ticks are in predetermined

locations and are not changed when the parent layout specifies axis
suboptions such as TICKVALUELIST=, TICKVALUESEQUENCE=,
and INCLUDERANGES=.

See “boolean ” on page 1339 for other Boolean values that you can use.

YBINSIZE=positive-number
specifies the size of bins along the Y role, in data units. The system determines the
NYBINS= and YBINSTART= values if they are not specified. The bins always span
the Y data range.

Default Determined by the system.

Interaction The YGAP= option is applied after this option.

YBINSTART=number
specifies the data value for the first bin of the Y role. The system determines the
NYBINS= and YBINSIZE= values if they are not specified. The bins always span
the Y data range.

Default Determined by the system.

Interaction The YVALUES= option specifies how this value is interpreted.

YBOUNDARY=UPPER | LOWER
specifies how an input value is counted when it lies on the endpoint of an Y bin. If
this option is set to UPPER, then the value is counted as one of the values in the
upper bin. Otherwise, it is counted in the lower bin.

Default UPPER
HEATMAP Statement 457

YENDLABELS=TRUE | FALSE
specifies whether the Y axis ticks and tick values are placed at the bin end-points or
at the bin mid-points.

Default FALSE. The axis ticks and tick values are placed at the bin mid-
points.

Interactions If YBINAXIS=FALSE, then this option is ignored.

This option is ignored when the Y column is non-numeric.

Tip The axis ticks and tick value placements are independent of the
YVALUES= option.

YGAP=number
specifies the amount of horizontal space on either side of each color-coded region in
the heat map.

Default 0

Interaction If the YBINSIZE= option is specified, the YGAP= value is subtracted

from the specified Y bin size.

YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the YBINSTART= value represents the midpoint, lower endpoint,
or upper endpoint of the bin.

Default MIDPOINTS

See YBINSTART= on page 456

YENDLABELS= on page 457

458 Chapter 6 • Plot Statements

Example: HEATMAP Statement

The following graph was generated by the “Example Program” on page 458.

Example Program
proc template;
define statgraph heatmap;
begingraph;
entrytitle "Vehicle Mileage By Curb Weight";
layout overlay /
xaxisopts=(label="Curb Weight (LBS)");
heatmap x=weight y=mpg_city / name="heatmap"
nybins=11 ybinstart=10 ybinsize=5
nxbins=11 xbinstart=2000 xbinsize=500;
continuouslegend "heatmap" / title="Count"
location=outside;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=heatmap;

run;
HEATMAPPARM Statement 459

HEATMAPPARM Statement
Creates a two-dimensional plot that represents the values of three variables. Generating an X, Y grid of
rectangles from the values of two independent variables, it colors the rectangles to represent the values of
a third variable, which can be a response variable or a group variable.
Requirements: The COLORGROUP= or the COLORRESPONSE= role must be specified.
The data for a parameterized heat map in a single-cell graph or in a cell in a
classification panel must have at least two bins for both the X and Y axes.
Otherwise, the heat map is not drawn.
Note: The data for a parameterized heat map should contain only one observation for each
X and Y value pair.
Tip: Starting with the third maintenance release of SAS 9.4, you can use subpixel
rendering with this statement. It is enabled by default. To disable subpixel rendering,
specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 33. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.

Syntax
HEATMAPPARM X=column | expression
Y=column | expression
COLORGROUP=column | discrete-attr-var | expression </option(s)>;
HEATMAPPARM X=column | expression
Y=column | expression
COLORRESPONSE=numeric-column | range-attr-var | expression </option(s)>;

Summary of Optional Arguments

Appearance options
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
DATATRANSPARENCY=number
specifies the degree of the transparency of the filled rectangles.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display outlined, colored rectangles or just colored
rectangles.
FILLATTRS=(TRANSPARENCY=number)
specifies the transparency of the area fill in the rectangles.
INCLUDEMISSINGCOLOR=TRUE | FALSE
specifies whether missing values of the color-group variable or of the color-
response variable are included in the plot.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the outlines of the filled rectangles.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.
460 Chapter 6 • Plot Statements

XGAP=number
specifies the amount of horizontal space on either side of each filled
rectangle.
XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input X values represent midpoints, lower endpoints, or
upper endpoints of the bins.
YGAP=number
specifies the amount of vertical space on either side of each filled rectangle.
YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input Y values represent midpoints, lower endpoints, or
upper endpoints of the bins.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
XBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for X-axis tick marks.
XBOUNDARIES=(numeric-list)
specifies the boundaries of the X-value bins.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
YBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for Y-axis tick marks.
YBOUNDARIES=(numeric-list)
specifies the boundaries of the Y-value bins.

Data option
DISCRETEX=TRUE | FALSE
specifies whether the X axis is discrete when X= specifies a numeric column.
DISCRETEY=TRUE | FALSE
specifies whether the Y axis is discrete when Y= specifies a numeric column.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a filled
rectangle.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
XENDLABELS=TRUE | FALSE
HEATMAPPARM Statement 461

specifies whether the axis ticks and value labels are drawn at the endpoints of
the bins or midpoints of the bins.
YENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the endpoints of
the bins or midpoints of the bins.

ODS options
URL=string-column
specifies an HTML page to display when a rectangle is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies the column for the X values.

Requirement If X= specifies a numeric column, then the DISCRETEX= option

must be set correctly for the X-axis type. When it is not set correctly,
the heat map is not drawn. If the X-axis type is discrete, then you
must specify DISCRETEX=TRUE in the HEATMAP statement.
Otherwise, DISCRETEX=FALSE must be in effect.

See DISCRETEX= on page 463

Y=column | expression
specifies the column for the Y values.

Requirement If Y= specifies a numeric column, then the DISCRETEY= option

must be set correctly for the Y-axis type. When it is not set correctly,
the heat map is not drawn. If the Y-axis type is discrete, then you
must specify DISCRETEY=TRUE in the HEATMAP statement.
Otherwise, DISCRETEY=FALSE must be in effect.

See DISCRETEY= on page 463

COLORGROUP=column | discrete-attr-var | expression

specifies a column or a discrete attribute variable that is used to discretely color the
regions in the heat map.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Requirement This argument is required when the response variable is of type

discrete.

See “DISCRETEATTRVAR Statement” on page 1297

462 Chapter 6 • Plot Statements

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies a numeric column or a range attribute variable that is used to color the
regions of the heat map.
range-attr-var
specifies a range attribute variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute variable specification must be a direct reference to

the attribute variable. It cannot be set by a dynamic variable.

When numeric-column or expression is specified, the colors for each region are
computed by mapping the values to a color ramp that is defined by the
COLORMODEL= option. When attr-variable is specified, the colors defined in the
associated RANGEATTRVAR or DISCRETEATTRVAR statement are used to color
the regions.

Requirement This argument is required when the response variable is of type

interval.

See “RANGEATTRVAR Statement” on page 1308

Optional Arguments
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to the
color model, use the REVERSECOLORMODEL= option.
HEATMAPPARM Statement 463

DATATRANSPARENCY=number
specifies the degree of the transparency of the filled rectangles.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISCRETEX=TRUE | FALSE
specifies whether the X axis is discrete when X= specifies a numeric column.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default FALSE

Requirements If X= specifies a numeric column and the X axis type is discrete,

then you must specify DISCRETEX=TRUE. Otherwise, the heat
map might not be drawn.

In SAS programs that were written before the third maintenance

release of SAS 9.4, if the HEATMAPPARM statement is used to
plot data on a discrete category axis, you must add the
DISCRETEX=TRUE option to the HEATMAPPARM statement.
Otherwise, the heat map is not drawn.

Interaction If X= specifies a character column, then this option is ignored, and

the X axis is considered to be discrete.

See X= on page 461

DISCRETEY=TRUE | FALSE
specifies whether the Y axis is discrete when Y= specifies a numeric column.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default FALSE

Requirements If Y= specifies a numeric column and the Y axis type is discrete,

then you must specify DISCRETEY=TRUE. Otherwise, the heat
map might not be drawn.

In SAS programs that were written before the third maintenance

release of SAS 9.4, if the HEATMAPPARM statement is used to
plot data on a discrete response axis, you must add the
DISCRETEY=TRUE option to the HEATMAPPARM statement.
Otherwise, the heat map is not drawn.

Interaction If Y= specifies a character column, then this option is ignored, and

the Y axis is considered to be discrete.

See Y= on page 461

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display outlined, colored rectangles or just colored rectangles.
STANDARD
displays colored rectangles
464 Chapter 6 • Plot Statements

ALL
displays outlined, colored rectangles
(display-options)
a space-separated list of one or more of options enclosed in parentheses.
Currently, only the OUTLINE option is supported, which displays outlines
around the filled rectangles (same as keyword ALL).

Default STANDARD

FILLATTRS=(TRANSPARENCY=number)
specifies the transparency of the area fill in the rectangles.

Default The DATATRANSPARENCY= option value

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction Only the TRANSPARENCY= suboption is honored. If a style element

or any other fill suboption is specified, then it is ignored by the
HEATMAPPARM statement.

Note The fill colors are determined by the COLORRESPONSE= or

COLORGROUP= or column.

INCLUDEMISSINGCOLOR=TRUE | FALSE
specifies whether missing values of the color-group variable or of the color-response
variable are included in the plot.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default TRUE

Tip The attributes of the missing color-group or color-response value are

determined by the GraphMissing style element unless a discrete attribute
map is in effect, the MISSING= system option changes the default missing
character, or a user-defined format is applied to the group value. In those
cases, the attributes of the missing group value are determined by a
GraphData1–GraphDataN style element instead of by the GraphMissing
style element.

See COLORGROUP= on page 461

COLORRESPONSE= on page 462

“boolean ” on page 1339 for other Boolean values that you can use.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the outlines of the filled rectangles.
HEATMAPPARM Statement 465

Default The ContrastColor and LineThickness attributes of the GraphOutlines

style element.

Interaction For this option to have any effect, outlines must be enabled by the ODS
style or the DISPLAY= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either the
ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

466 Chapter 6 • Plot Statements

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, COLORGROUP=, and
COLORRESPONSE=.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a filled
rectangle. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the heat map can be
specified along with roles that do.
(role-list)
an ordered, space-separated list of unique HEATMAPPARM and user-defined
roles. HEATMAPPARM roles include X and Y.

Example The following example displays data tips for the columns X and PCT.
The Pct column is not assigned to any pre-defined HEATMAPPARM
role, so it must first be assigned a role:
ROLENAME=(TIP1=PCT)
TIP=(X TIP1)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X and Y, and COLORGROUP= or
COLORRESPONSE=.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP variables can be controlled with
the TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.
HEATMAPPARM Statement 467

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a rectangle is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
rectangle that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate a plot with selectable rectangles, you must include an

ODS GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some X and Y pairs. In that case, no
action is taken when the corresponding rectangle is selected.

The URL value can be the same for any X and Y pairs. In that case,
the same action is taken when the rectangle for those X and Y pairs is
selected.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X
468 Chapter 6 • Plot Statements

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

XBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for X-axis tick marks. When this option is
FALSE, a standard axis is used, ignoring bin boundaries and midpoints.

Default TRUE

Interactions When this option is TRUE, the XENDLABELS= option determines

how the axis ticks and value labels are displayed. When this option is
FALSE, the XENDLABELS= option is ignored.

When this option is TRUE, the axis ticks are in predetermined

locations and are not changed when the parent layout specifies axis
suboptions such as TICKVALUELIST=, TICKVALUESEQUENCE=,
and INCLUDERANGES=.

See “boolean ” on page 1339 for other Boolean values that you can use.

XBOUNDARIES=(numeric-list)
specifies the boundaries of the X-value bins. The boundaries are specified as a space-
separated list of values enclosed in parentheses. The keywords MIN and MAX can
be used as one of the values in the list of boundaries. Keywords MIN and MAX
indicate the minimum and maximum data values for the X variable.

Interaction This option is ignored if the X values are not numeric.

Tip This option can be used to specify unequal bins.

Example xboundaries=(MIN 20 200 250 MAX)

XENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the endpoints of the
bins or midpoints of the bins.

Default FALSE. The axis ticks and values labels are drawn at the bin
midpoints.

Interactions If this option is set to FALSE, then the axis ticks and value labels are
drawn at the bin midpoints, regardless of whether the XVALUES=
option identifies the X data as endpoint values or midpoint values.

This option is ignored if the X values are not numeric or if

XBINAXIS=FALSE.

See “boolean ” on page 1339 for other Boolean values that you can use.

XGAP=number
specifies the amount of horizontal space on either side of each filled rectangle.

Default 0

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input X values represent midpoints, lower endpoints, or upper
endpoints of the bins.
HEATMAPPARM Statement 469

Default MIDPOINTS

See XENDLABELS=

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for Y-axis tick marks. When this option is
FALSE, a standard axis is used, ignoring bin boundaries and midpoints.

Default TRUE

Interactions When this option is TRUE, the YENDLABELS= option determines

how the axis ticks and value labels are displayed. When this option is
FALSE, the YENDLABELS= option is ignored.

When this option is TRUE, the axis ticks are in predetermined

locations and are not changed when the parent layout specifies axis
suboptions such as TICKVALUELIST=, TICKVALUESEQUENCE=,
and INCLUDERANGES=.

See “boolean ” on page 1339 for other Boolean values that you can use.

YBOUNDARIES=(numeric-list)
specifies the boundaries of the Y-value bins. The boundaries are specified as a space-
separated list of values enclosed in parentheses. The keywords MIN and MAX can
be used as one of the values in the list of boundaries. Keywords MIN and MAX
indicate the minimum and maximum data values for the Y variable. Example:
yboundaries=(MIN 20 200 250 MAX)

Interaction This option is ignored if the Y values are not numeric.

Tip This option can be used to specify unequal bins.

YENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the endpoints of the
bins or midpoints of the bins.

Default FALSE. The axis ticks and values labels are drawn at the bin
midpoints.

Interactions If this option is set to FALSE, then the axis ticks and value labels are
drawn at the bin midpoints, regardless of whether the YVALUES=
option identifies the Y data as endpoint values or midpoint values.

This option is ignored if the Y values are not numeric or if

YBINAXIS=FALSE.

See “boolean ” on page 1339 for other Boolean values that you can use.
470 Chapter 6 • Plot Statements

YGAP=number
specifies the amount of vertical space on either side of each filled rectangle.

Default 0

YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input Y values represent midpoints, lower endpoints, or upper
endpoints of the bins.

Default MIDPOINTS

See YENDLABELS=

Details
A heat map is useful for visualizing the magnitude of a response variable in relation to
two independent variables. For example, in molecular biology, heat maps can be used to
track the expression of genes across multiple sample studies. In the HEATMAPPARM
statement, you specify the independent variables in the X and Y arguments. For a
response variable that has discrete values, use the COLORGROUP argument, or for a
response variable that has interval values, use the COLORRESPONSE argument. For
interval response variables, you can use the COLORMODEL= option to define the color
ramp that is used to fill the rectangles.

Example: HEATMAPPARM Statement

The following graph was generated by the “Example Program” on page 471:
HIGHLOWPLOT Statement 471

Example Program
proc template;
define statgraph heatmapparm;
begingraph;
layout overlay;
heatmapparm x=height y=weight colorresponse=count /
name="heatmapparm" xbinaxis=false ybinaxis=false;
continuouslegend "heatmapparm" / location=outside valign=bottom;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.gridded template=heatmapparm;

run;

HIGHLOWPLOT Statement
Creates a display of floating vertical or horizontal lines or bars that connect the minimum and maximum
response values for each value of a categorical variable.
Requirements: Either the X= or Y= argument must be specified, but you cannot specify both on the
same HIGHLOWPLOT statement.
The HIGH= and LOW= arguments are required.
Note: Specifying the X= option creates a vertical high-low chart, which would typically be
used in the financial industry to plot stock values over time. Specifying the Y= option
creates a horizontal high-low chart, which would typically be used in the Health and
Life Sciences industry to display over time the duration of adverse events or of
adverse reactions to medication.
Tips: For charts that have a large number of bars that are very close together, slight
variations in spacing that normally occur due to integer rounding can become more
obvious. Subpixel rendering provides more precise bar spacing in that case. In the
second maintenance release of SAS 9.4 and in earlier releases, specify
SUBPIXEL=ON in the BEGINGRAPH statement to enable subpixel rendering. See
SUBPIXEL= on page 33. Starting with the third maintenance release of SAS 9.4,
subpixel rendering is enabled by default.
To disable subpixel rendering in the third maintenance release of SAS 9.4 and in
later releases, specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an
ODS GRAPHICS statement. For information about the BEGINGRAPH statement
SUBPIXEL= option, see SUBPIXEL= on page 33. For information about the ODS
GRAPHICS statement SUBPIXEL= option, see “ODS GRAPHICS Statement” in
SAS ODS Graphics: Procedures Guide.

Syntax
HIGHLOWPLOT X=column | expression
LOW=numeric-column | expression
HIGH=numeric-column | expression </option(s)>;
HIGHLOWPLOT Y=column | expression
LOW=numeric-column | expression
HIGH=numeric-column | expression </option(s)>;
472 Chapter 6 • Plot Statements

Summary of Optional Arguments

Appearance options
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.
CLIPCAP=TRUE | FALSE
specifies whether a special clip cap is displayed to indicate where clipping
occurred.
CLIPCAPSHAPE=DEFAULT | BARBEDARROW | CLIPPEDARROW |
CLOSEDARROW | FILLEDARROW | OPENARROW | SERIF
specifies the shape of the arrowhead on the clipped end of a line or bar when
CLIPCAP=TRUE.
CLOSE=numeric-column | expression
specifies a column or expressions whose values are used to display a closing-
value indicator.
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group
clusters as a fraction of the minimum interval between adjacent data values.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the bar or line
colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars or lines of a high-low chart.
DATATRANSPARENCY=number
specifies the degree of the transparency of the colored regions, and the high-
end and low-end labels.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display outlined colored regions or just colored regions.
ENDCAPDISPLAYPOLICY=AUTO | ALWAYS
specifies the policy for displaying end caps when end caps are present.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the bars.
HIGHCAP=column | NONE | SERIF | BARBEDARROW | FILLEDARROW |
OPENARROW | CLOSEDARROW
specifies the type of cap used at the high end of the bar or line.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) or bar
attributes (fill and outline) to one of the GraphData1–GraphDataN style
elements.
INTERVALBARWIDTH=dimension
specifies the width of the floating bars.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the floating plot lines.
LOWCAP=column | NONE | SERIF | BARBEDARROW | FILLEDARROW |
OPENARROW | CLOSEDARROW
specifies the type of cap used at the low end of the bar or line.
OPEN=numeric-column | expression
HIGHLOWPLOT Statement 473

specifies a column or expressions whose values are used to display an

opening-value indicator.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the outlines of the filled regions.
TYPE=LINE | BAR
specifies whether data values should be represented by bars or lines.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar
or line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
HIGHLABEL=column | expression
specifies the label to display at the high end of the bar or line.
LABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the labels for the bars or lines.
LOWLABEL=column | expression
specifies the label to appear at the low end of each floating bar or line.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all bars or lines from the category midpoints
when graphing multiple response variables side by side on a common axis.
GROUP=column | discrete-attr-var | expression
creates a distinct set of floating bars or lines for each unique group value in
the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped bars or lines are overlaid or clustered around the
category midpoints.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
474 Chapter 6 • Plot Statements

specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
Either the X= or Y= argument must be specified, but you cannot specify both on the
same HIGHLOWPLOT statement. The HIGH= and LOW= arguments are required.
X=column | expression
specifies a column or expression representing the values in a vertical high-low chart.
The values can be character or numeric.
Y=column | expression
specifies a column or expression representing the values in a horizontal high-low
chart. The values can be character or numeric.
LOW=numeric-column | expression
specifies a column or expression representing the values of the lower end of the
floating bar or line.
HIGH=numeric-column | expression
specifies a column or expression representing the values of the higher end of the
floating bar or line.

Optional Arguments
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.

Default 0.85

Range 0.1–1, where 0.1 is the narrowest and 1 is the widest

Restriction This option has an effect only when TYPE= BAR.

Interaction Prior to the third maintenance release of SAS 9.4, this option is ignored
for an interval high-low plot, and the bar width is controlled by the
INTERVALBARWIDTH= option. Starting with the third maintenance
release of SAS 9.4, this option is honored for an interval high-low plot
but it can be overridden by the INTERVALBARWIDTH= option.

Notes This option is needed only to change the default behavior.

By default, the bar width automatically adjusts based on the number of

bars to be displayed and the wall width.

Tip To remove any inter-bar gap, set BARWIDTH=1.

See DISCRETEOFFSET= option for examples of using this option.

CLIPCAP=TRUE | FALSE
specifies whether a special clip cap is displayed to indicate where clipping occurred.
When the VIEWMIN= and VIEWMAX= axis options are specified for an axis and a
data value exceeds the specified axis range, the bar for that value is clipped. In that
HIGHLOWPLOT Statement 475

case, when CLIPCAP=TRUE, a special clip cap is displayed at the clipped end of the
bar. If the bar already has a high or low cap, then it is replaced by the clip cap.
For vertical bars, the clip cap is added to the end of the bar that is clipped by the Y
axis range. The default clip cap is a vertical clipped arrowhead that points toward the
clip edge ( or ). For horizontal bars, the cap is added to the end of the bar that is
clipped by the X axis range. The default clip cap is a horizontal clipped arrowhead
that points toward the clip edge ( or ). If an entire bar is clipped, then a clip cap
is displayed at the high or low side where the bar was clipped.
The following figure shows a side-by-side comparison of a vertical line high-low
chart with no clipping and with clipping when CLIPCAP=TRUE.

The first graph shows the default case in which no clipping occurs. The second graph
shows the case in which VIEWMIN= and VIEWMAX= specify a range that causes
clipping on the Y axis. In the second graph, the clip caps indicate clipping for each
category line as follows:

A the high end of the line is clipped.

B the low end of the line is clipped.
C both ends of the line are clipped.
D the entire line is clipped at the low end.
E the entire line is clipped at the high end.

Default FALSE

Interaction Clip caps appear only when CLIPCAP=TRUE and the data values
exceed the axis range that is specified by the VIEWMIN= and
VIEWMAX= options. When the VIEWMIN= and VIEWMAX=
options are not specified, the axis range is adjusted to accommodate the
data values and clipping does not occur.

Note If the HIGHLABEL= and LOWLABEL= options are in effect and the
bar is clipped, then the label value at the clipped end is drawn at the tip
of the clip cap. If the entire bar is clipped, then the labels are not
shown.
476 Chapter 6 • Plot Statements

Tip Use the CLIPCAPSHAPE= option to specify a different clip-cap

arrowhead.

See “boolean ” on page 1339 for other Boolean values that you can use.

CLIPCAPSHAPE=DEFAULT | BARBEDARROW | CLIPPEDARROW |

CLOSEDARROW | FILLEDARROW | OPENARROW | SERIF
specifies the shape of the arrowhead on the clipped end of a line or bar when
CLIPCAP=TRUE. The following figure shows each of the clip-cap arrowhead
shapes for vertical lines and bars that are clipped at both ends.

In the first maintenance release of SAS 9.4 and earlier releases, this option specifies
the arrowhead shape only for the clipped end of clipped lines (TYPE=LINE). For
bars (TYPE=BAR), this option is ignored, and CLIPPEDARROW is always used as
the arrowhead shape for the clipped end of clipped bars.
Starting with the second maintenance release of SAS 9.4, this option specifies the
arrowhead shape for the clipped end of clipped lines or clipped bars. For bars, you
can specify the CLIPPEDARROW shape for filled and unfilled bars. When any
value other than CLIPPEDARROW or DEFAULT is specified for bars, the
FILLEDARROW shape is used for filled bars, and the CLOSEDARROW shape is
used for unfilled bars.

Default DEFAULT, which is the same as CLIPPEDARROW

Interaction This option is ignored when CLIPCAP=FALSE.

CLOSE=numeric-column | expression
specifies a column or expressions whose values are used to display a closing-value
indicator. This option is typically used when TYPE=LINE, but it can be used when
TYPE=BAR. For vertical high-low charts, the value is represented by a short
horizontal line extending from the side that displays the higher X values. For
horizontal high-low charts, the value is represented by a short vertical line extending
from the side that displays the higher Y values.
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters as a
fraction of the minimum interval between adjacent data values.

Default 0.85
HIGHLOWPLOT Statement 477

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.

Interaction For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Defaults For lines or for outline-only bars, the ThreeColorAltRamp style

element.

For bars with fill, the ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the bar or line colors.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.
478 Chapter 6 • Plot Statements

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column has
multiple values for a single GROUP value, unexpected results might
occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map.

When only the outlines are displayed, this option overrides suboption
COLOR= in the OUTLINEATTRS= option and varies the outline
color according to the color gradient or the attribute map.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

For a numeric column or expression, the ThreeColorRamp style

element defines the fill color gradient, and the ThreeColorAltRamp
style element defines the outline color gradient.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars or lines of a high-low chart. The
following figure shows a high-low chart that contains a filled bar and a line with
each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the GraphSkins:DataSkin
style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.
HIGHLOWPLOT Statement 479

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

When TYPE=BAR, the DISPLAY= option in effect must include

FILL in order for the DATASKIN= option to have any effect.

For filled bars, the skin appearance is based on the FILLATTRS=

option color specification.

DATATRANSPARENCY=number
specifies the degree of the transparency of the colored regions, and the high-end and
low-end labels.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS option can be used to set transparency for just the colored
regions. You can combine this option with FILLATTRS= to set one
transparency for the outlines but a different transparency for the region
fills . Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEOFFSET=number
specifies an amount to offset all bars or lines from the category midpoints when
graphing multiple response variables side by side on a common axis.

Default 0 (no offset, all bars or lines are centered on the category midpoints)

Range –0.5 to +0.5, where 0.5 represents half the distance between category
ticks. Normally, a positive offset is to the right for a vertical plot and up
for a horizontal plot. (If the layout's axis options set
REVERSE=TRUE, then the offset direction is also reversed.)

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

See “About the DISCRETEOFFSET= Option” on page 490

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display outlined colored regions or just colored regions.
STANDARD
displays outlined, colored regions
ALL
displays outlined, colored regions
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:
480 Chapter 6 • Plot Statements

FILL displays filled (colored) regions

OUTLINE displays outlined regions

Default STANDARD

Interaction For this option to take effect, TYPE=BAR must also be specified.

ENDCAPDISPLAYPOLICY=AUTO | ALWAYS
specifies the policy for displaying end caps when end caps are present.
AUTO
draws the end caps only for the elements with a low-to-high range that is large
enough to accommodate the end caps. Draws just the bar or line for the rest.
ALWAYS
always draws the end caps.

Note When the low-to-high range is not large enough for the end caps, the end
caps might overlap. The data ranges that are smaller than the end-cap size
might not be resolvable beyond the rendered size of the end caps.

Default AUTO

Interactions This option is honored only when the option TYPE=LINE is in effect.

This option is ignored if the LOWCAP= or HIGHCAP= option is not

set or is effectively set to NONE.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the bars.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

For grouped data, the Color attribute of GraphData1–GraphDataN

style elements.

Interactions For this option to take effect, TYPE=BAR must also be specified, and
the DISPLAY= option must allow the fill to be displayed.

When COLORRESPONSE= is in effect, the FILLATTRS= suboption

COLOR= is ignored, and the bar fill colors vary according to the
gradient.

Tip The DATATRANSPARENCY= option sets the transparency for the

colored regions and the outlines around them. You can combine this
option with DATATRANSPARENCY= to set one transparency for the
outlines but a different transparency for the colored regions. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

GROUP=column | discrete-attr-var | expression

creates a distinct set of floating bars or lines for each unique group value in the
specified column.
HIGHLOWPLOT Statement 481

discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Defaults For filled bars, each distinct group value is represented in the plot by a
different fill color. The fill colors are defined by the Color attribute of
the GraphData1–GraphDataN and GraphMissing style elements.

For floating lines and outlined bars, each distinct group value is
represented in the plot by a different line color and line pattern. The
line colors and line patterns are defined by the ContrastColor and
LineStyle attributes of the GraphData1–GraphDataN and
GraphMissing style elements.

Interactions If the X or Y role is discrete, then the bars or lines representing the
group can be drawn in a cluster or overlaid, based on the setting for
the GROUPDISPLAY= option. However, if the X or Y role is interval,
then the lines representing the group are overlaid and the
GROUPDISPLAY= option is ignored.

By default, the group values are mapped in the order of the data. Use
the GROUPORDER= option to control the sorting order of the
grouped bar segments. Use the INDEX= option to alter the default
sequence of colors and line patterns.

The INCLUDEMISSINGGROUP= option determines whether

missing group values are considered a distinct group value.

When both the GROUP= and COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.

Tip The representations that are used to identify the groups can be
overridden. For example, if each distinct group value is represented by
a different line pattern, you can use the
LINEATTRS=(PATTERN=pattern) specification to assign the same
line pattern to all the plot lines. In that case, the line color denotes
group values. Likewise, you can use LINEATTRS=(COLOR=color)
to assign the same color to all lines, letting line pattern denote group
values.

See “DISCRETEATTRVAR Statement” on page 1297

GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped bars or lines are overlaid or clustered around the category
midpoints.
OVERLAY
centers the bars or lines for matching category values on the midpoints. The bars
or lines in each set of group values are superimposed on each other.
CLUSTER
clusters the bars or lines for matching category values around the midpoints.
Each cluster of group values is centered at the midpoint for the category.
482 Chapter 6 • Plot Statements

The following example shows the effect of clustering the lines in a stock report when
the category values are grouped into a single response variable. Note that if your
category values are not grouped in the same column but are stored in separate
columns, then you can get this same effect by using the DISCRETEOFFSET=
option.
layout overlay /
yaxisopts=(label="Stock Value");
highlowplot x=month high=high low=low /
close=close open=open
legendlabel="Stock" name="cluster"
group=stock groupdisplay=cluster
lineattrs=(pattern=solid);
discretelegend "cluster" / title="Stock"
location=inside halign=right valign=top;
endlayout;

Default OVERLAY

Interaction For this option to take effect, the GROUP= option must also be
specified.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
HIGHLOWPLOT Statement 483

DESCENDING
orders the groups within a category in descending order.

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

This option is ignored when GROUPDISPLAY=OVERLAY.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.

HIGHCAP=column | NONE | SERIF | BARBEDARROW | FILLEDARROW |

OPENARROW | CLOSEDARROW
specifies the type of cap used at the high end of the bar or line. All of the keywords
can be specified for any high-low chart. The effect of each keyword depends on the
setting for the TYPE= and DISPLAY= options as follows:
• When TYPE=BAR and DISPLAY= includes FILL, FILLEDARROW is used for
all settings other than NONE.
• When TYPE=BAR and DISPLAY= does not include FILL, CLOSEDARROW is
used for all settings other than NONE.
• When TYPE=LINE and CLOSEDARROW is specified, FILLEDARROW is
used instead.
The following figure shows the effect of each cap value on horizontal lines, filled
bars, and unfilled bars.
Figure 6.1 Horizontal High and Low Cap Shapes for Lines, Filled Bars, and Unfilled Bars
484 Chapter 6 • Plot Statements

Default NONE

Interaction When TYPE=BAR, the caps are drawn to fit within the bar width. The
width of the bar itself is reduced.

Note If the length of the high-low element is smaller than the cap, then the
cap is not drawn.

HIGHLABEL=column | expression
specifies the label to display at the high end of the bar or line.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) or bar attributes
(fill and outline) to one of the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped
according to the setting of the GROUPORDER= option.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.
HIGHLOWPLOT Statement 485

INTERVALBARWIDTH=dimension
specifies the width of the floating bars.

Defaults Prior to the third maintenance release of SAS 9.4, 85% of the smallest
interval between any two bars for the given plot

Starting with the third maintenance release of SAS 9.4, the

BARWIDTH= option setting

Restriction For this option to take effect, TYPE=BAR must be set, and the
independent variable must be of type interval.

Interaction Prior to the third maintenance release of SAS 9.4, this option controls
the bar width for a high-low plot. Starting with the third maintenance
release of SAS 9.4, this option overrides the BARWIDTH= option.

See “dimension” on page 1340

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the labels for the bars or lines.

Defaults In the second maintenance release of SAS 9.4 and earlier releases, the
GraphDataText style element.

Starting with the second maintenance release of SAS 9.4, the

GraphDataText style element is the default for non-grouped data. For
grouped data, the data label color is determined by the ContrastColor
attribute of the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element.

“Text Options” on page 1351 for available text-options.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the floating plot lines.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor, LineStyle, and LineThickness

attributes of the GraphData1–GraphDataN style elements.

Interaction For this option to have any effect, TYPE= LINE must also be specified.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

LOWCAP=column | NONE | SERIF | BARBEDARROW | FILLEDARROW |

OPENARROW | CLOSEDARROW
specifies the type of cap used at the low end of the bar or line. All of the keywords
can be specified for any high-low chart. The effect of each keyword depends on the
setting for the TYPE= and DISPLAY= options as follows:
• When TYPE=BAR and DISPLAY= includes FILL, FILLEDARROW is used for
all settings other than NONE.
486 Chapter 6 • Plot Statements

• When TYPE=BAR and DISPLAY= does not include FILL, CLOSEDARROW is

used for all settings other than NONE.
• When TYPE=LINE and CLOSEDARROW is specified, FILLEDARROW is
used instead.
Figure 6.1 on page 483 shows the effect of each cap value on horizontal lines, filled
bars, and unfilled bars.

Default NONE

Interaction When TYPE=BAR, the caps are drawn to fit within the bar width. The
width of the bar itself is reduced.

Note If the length of the high-low element is smaller than the cap, then the
cap is not drawn.

LOWLABEL=column | expression
specifies the label to appear at the low end of each floating bar or line.
NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

OPEN=numeric-column | expression
specifies a column or expressions whose values are used to display an opening-value
indicator. This option is typically used when TYPE=LINE, but it can be used when
TYPE=BAR. For vertical high-low charts, the value is represented by a short
horizontal line extending from the side that displays the lower X values. For
horizontal high-low charts, the value is represented by a short vertical line extending
from the side that displays the lower Y values.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the outlines of the filled regions.

Defaults For non-grouped data, the ContrastColor, LineThickness, and

LineStyle attributes of the GraphOutlines style element.

For grouped data when FILL is displayed, the LineThickness and

LineStyle attributes of the GraphOutlines style element, and the
ContrastColor attribute of the GraphData1–GraphDataN style
elements.

For grouped data when FILL is not displayed, the LineThickness

attribute of the GraphOutlines style element, and the ContrastColor
and LineStyle attributes of the GraphData1–GraphDataN style
elements.

Interactions For this option to have any effect, TYPE=BAR must be specified, and
outlines must be enabled by the ODS style or the DISPLAY= option.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the bar outline colors vary according to the gradient.
HIGHLOWPLOT Statement 487

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, LOW, HIGH, OPEN, and CLOSE.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar or line.
If this option is used, then the information specified replaces all of the information
that is displayed by default. You can specify roles for columns that do not contribute
to the bar chart along with roles that do.
(role-list)
an ordered, space-separated list of unique HIGHLOWPLOT and user-defined
roles. HIGHLOWPLOT roles include X, Y, LOW, HIGH, OPEN, CLOSE,
GROUP, and COLORRESPONSE.

Note The COLORRESPONSE role is valid starting with the third

maintenance release of SAS 9.4.
488 Chapter 6 • Plot Statements

Example The following example displays data tips for the columns X and Pct.
The Pct column is not assigned to any pre-defined HIGHLOWPLOT
role, so it must first be assigned a role.
ROLENAME=(TIP1=PCT)
TIP=(X TIP1)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X or Y and GROUP.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement with the IMAGEMAP option specified,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.
HIGHLOWPLOT Statement 489

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TYPE=LINE | BAR
specifies whether data values should be represented by bars or lines.

BAR uses fill and outline attributes.

LINE uses line attributes.

Default LINE

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interactions When the Y= argument is specified, the HIGH= and LOW=

arguments are mapped to the axis that is specified on this option (X or
X2).

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interactions When the X= argument is specified, the HIGH= and LOW=

arguments are mapped to the axis that is specified on this option (Y or
Y2).

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details

Statement Description
A high-low chart specifies that floating vertical or horizontal lines or bars connect the
minimum and maximum response values for each value of a categorical variable. The
data should have at least two response values for every category value. Otherwise, the
single value is displayed without the connecting line or bar. In the statement syntax, the
X and Y arguments can specify a column with character or numeric values. The LOW
and HIGH arguments must specify a numeric column.
The HIGHLOWPLOT statement can be used to create a plot showing the high and low
response values for observations along a time axis. The independent variable can be
interval or discrete. The HIGHLOWPLOT has two common uses:
• One typical use of a high-low chart is in the financial industry to plot stock values by
day. The OPEN= and CLOSE= options are typically used in the plot, and the TYPE=
490 Chapter 6 • Plot Statements

option is set to LINE (the default), as illustrated in “Example 1: Vertical High-low

Chart” on page 491 .
• Another typical use of a high-low chart is in the Health and Life Sciences industry to
display over time the duration of adverse events or of adverse reactions to
medication. The HIGHCAP= and LOWCAP= options are typically used in the plot,
and the TYPE= option is set to BAR, as illustrated in “Example 2: Horizontal High-
low Chart” on page 492 .
For both vertical and horizontal high-low charts, the data values can be grouped using
the GROUP= option. For grouped data, the data values are not summarized. Each
observation is plotted independently, and the line or bar segment for each category value
can use different display characteristics.

About the DISCRETEOFFSET= Option

This feature is useful for graphing multiple response variables side by side on a common
axis. By default within an overlay-type layout, if multiple HIGHLOWPLOT statements
are used with different response variables, then the bars or lines for matching category
values are centered on the midpoints and superimposed on each other. In those cases,
you can make it easier to distinguish among superimposed bars or lines by assigning a
different width for them in each HIGHLOWPLOT statement in the overlay.
If you prefer to avoid superimposed bars or lines, then you can assign a different offset
to each HIGHLOWPLOT statement. If desired, you can adjust the width of the bar or
line in conjunction with DISCRETEOFFSET= to create narrower bars that require less
width within the plot area.
The following example defines offsets for the lines in a stock report:
layout overlay / cycleattrs=true
yaxisopts=(label="Stock Value");
highlowplot x=month high=a_high low=a_low /
close=a_close open=a_open
legendlabel="A" name="ahighlow" lineattrs=(pattern=solid)
discreteoffset=-0.2 ;
highlowplot x=month high=b_high low=b_low /
close=b_close open=b_open
legendlabel="B" name="bhighlow" lineattrs=(pattern=solid)
discreteoffset=0 ;
highlowplot x=month high=c_high low=c_low /
close=c_close open=c_open
legendlabel="C" name="chighlow" lineattrs=(pattern=solid)
discreteoffset=+0.2 ;
discretelegend "ahighlow" "bhighlow" "chighlow" / title="Stock"
location=inside halign=right valign=top;
endlayout;
Example 1: Vertical High-low Chart 491

Examples

Example 1: Vertical High-low Chart

The following vertical high-low chart was generated by “Example Program” on page
491:

Example Program
proc template;
define statgraph highlow;
begingraph;
layout overlay /
492 Chapter 6 • Plot Statements

yaxisopts=(griddisplay=on label="Stock Value");

highlowplot x=date high=high low=low /
open=open close=close;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.stocks template=highlow;

where date >= '01JAN03'd and stock="IBM";
run;

Example 2: Horizontal High-low Chart

The following horizontal high-low chart was generated by “Example Program” on page
492 :

Example Program
data highlowbar;
length cap $ 12;
input drug $ 1-10 low high cap $;
datalines;
Drug A 10 20 NONE
Drug A 30 60 FILLEDARROW
Drug B 20 35 NONE
Drug B 50 75 FILLEDARROW
Drug C 30 90 FILLEDARROW
;
proc template;
define statgraph highlowbar;
begingraph;
entrytitle 'Medications Plot';
layout overlay /
HISTOGRAM Statement 493

xaxisopts=(
griddisplay=on
linearopts=(viewmin=0 viewmax=100)
display=(line ticks tickvalues))
yaxisopts=(
griddisplay=on
display=(line ticks tickvalues));
highlowplot y=drug high=high low=low /
group=drug outlineattrs=(pattern=solid)
datatransparency=0.4
type=bar barwidth=0.4
highcap=cap lowcap=cap;
endlayout;
endgraph;
end;

proc sgrender data=highlowbar template=highlowbar;

run;

HISTOGRAM Statement
Creates a univariate histogram computed from input data.
Note: When a histogram is placed on a log axis, binning is done on a linear scale, but the
bins are drawn on the log scale. As a result, the bins might have different widths
along the log axis.

Syntax
HISTOGRAM numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
BINSTART=number
specifies the X coordinate of the first bin.
BINWIDTH=positive-number
specifies the bin width.
BOUNDARY=UPPER | LOWER
specifies how a boundary is counted when it lies on the endpoint of a bin.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fills and bar outlines.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display outlined bars, filled bars, or outlined and filled
bars.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the bars.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled bar area.
494 Chapter 6 • Plot Statements

FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.
NBINS=positive-integer
specifies the number of bins.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line properties of the bar outlines.
XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input X values represent midpoints, lower endpoints, or
upper endpoints of the bins.

Axes options
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
SCALE=PERCENT | COUNT | PROPORTION | DENSITY
specifies the scale for the Y axis.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABELTYPE=NONE | AUTO | COUNT | DENSITY | PERCENT |
PROPORTION
specifies the statistic to display at the end of each bar.
ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin
endpoints (TRUE) or at the bin midpoints (FALSE).
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate bar segment or bar for each unique group value in the
specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.
HISTOGRAM Statement 495

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation
that is read.
WEIGHT=numeric-column | expression
specifies a column that contains a bin-width calculation a priori weight for
each observation of the input data object.

Required Argument
numeric-column | expression
specifies a column that contains numeric values, or a dynamic variable that refers to
such a column.

Optional Arguments
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.
TRUE
specifies that the ENDLABELS= option determines how the axis ticks and value
labels are displayed.
FALSE
specifies that standard axes are used. Bin boundaries and midpoints that are set
by the ENDLABELS= option are ignored.

Default TRUE

Interactions This option is ignored when this histogram is not the primary plot. For
more information about primary plots, see “When Plots Share Data
and a Common Axis” on page 880.

BINAXIS=TRUE is ignored when the histogram is placed on a log

axis.

When this option is set to TRUE, some X-axis options that are set on
the parent layout might not apply, such as INTEGER=,
TICKVALUELIST=, TICKVALUESEQUENCE=, and
INCLUDERANGES=.

See “boolean ” on page 1339 for other Boolean values that you can use.

BINSTART=number
specifies the X coordinate of the first bin. This option can be used with the
BINWIDTH= and NBINS= options to specify the bins.

Default Determined by the system.

Interaction This option is ignored if the specified number is greater than the
minimum data value. In that case, the default starting bin is used
instead.
496 Chapter 6 • Plot Statements

Note If the BINSTART= value is less than the minimum data value, then the
lower end of the histogram might be padded with zero-height bins in
order to accommodate the BINSTART= value.

See XVALUES=

BINWIDTH=positive-number
specifies the bin width. This option can be used with the BINSTART= and NBINS=
options to specify the bins.

Default Determined by the system.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of bins is limited to approximately 10,000. If the number of
bins computed from the data and the BINWIDTH= value exceeds
10,000, SAS computes a new bin-width value that yields approximately
10,000 bins. A warning of the change is written to the SAS log in that
case.

Interaction This option is ignored when the NBINS= option is also specified (with
or without the BINSTART= option) and the resulting data range does
not completely span the unbinned input data range. In that case, the
NBINS= option is honored and the default bin width is used instead.

Note When BINSTART=, NBINS=, or both are specified with this option
and the resulting data range is greater than the unbinned input data
range, the histogram might be padded with zero-height bins in order to
accommodate the BINSTART=, NBINS=, and BINWIDTH= option
values.

BOUNDARY=UPPER | LOWER
specifies how a boundary is counted when it lies on the endpoint of a bin. If this
option is set to UPPER, then the value is counted as one of the values in the upper
bin (the bin to the right). Otherwise, it is counted in the lower bin.

Default UPPER

DATALABELTYPE=NONE | AUTO | COUNT | DENSITY | PERCENT |

PROPORTION
specifies the statistic to display at the end of each bar.
NONE
suppresses the data labels.
AUTO
uses the SCALE= option value. By default, SCALE=PERCENT.
COUNT | DENSITY | PERCENT | PROPORTION
specifies that the count, density, percentage, or proportion statistic is to be
displayed at the end of each bar.

Default NONE

Interaction When DATALABELTYPE=AUTO, the SCALE= option determines the

statistic that is displayed at the end of each bar.
HISTOGRAM Statement 497

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows
histogram bars with each of the skins applied.

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

When FILLTYPE=GRADIENT is in effect, DATASKIN=SHEEN is

ignored. In that case, use one of the other skins.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fills and bar outlines.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the bar
fills. You can combine this option with FILLATTRS= to set one
transparency for the bar outlines but a different transparency for the bar
fills. Example:
498 Chapter 6 • Plot Statements

datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display outlined bars, filled bars, or outlined and filled bars.
STANDARD
displays outlined, filled bars. (Currently, the same as ALL.)
ALL
displays outlined, filled bars.
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:
OUTLINE
displays outlined bars.
FILL
displays bars with a color fill.
FILLPATTERN
displays bars with a patterned fill.

Note This option is valid in the first maintenance release of SAS 9.4 and
later releases.

Default The GraphHistogram:DisplayOpts style reference.

Tip Use the OUTLINEATTRS= , FILLATTRS= , and FILLPATTERNATTRS=

options to control the appearance of the bars.

ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin endpoints
(TRUE) or at the bin midpoints (FALSE).

Default FALSE.

Interactions This option is ignored if this plot is not the primary plot in the parent
layout. For more information, see the PRIMARY= option.

This option is ignored if BINAXIS= FALSE. By default,

BINAXIS=TRUE.

If the TICKS= suboption is specified in the XAXISOPTS= option,

then this option is ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the bars.

Defaults In the first maintenance release of SAS 9.4 and earlier releases, the
Color attribute of the GraphDataDefault style element

Starting with the second maintenance release of SAS 9.4, the Color
attribute of the GraphDataDefault style element for non-grouped data or
the GraphData1–GraphDataN style elements for grouped data.
HISTOGRAM Statement 499

Interaction For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

Tip The DATATRANSPARENCY= option sets the transparency for bar fills
and bar outlines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines but
a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled bar area.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
style-element
specifies the name of a style element. You can specify only one of the elements
GraphData1–GraphDataN.

Restriction The only styles that are delivered by SAS that support fill patterns
are JOURNAL2, JOURNAL3, and MONOCHROMEPRINTER. If
any other such style is in effect and this option uses style-element in
its specification, then this option is ignored.

(fill-pattern-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:
COLOR=color | style-reference
specifies a color to use for the bar-fill-pattern lines.
PATTERN=line-pattern
specifies a line pattern to use for the bar fill.
To specify a line-pattern, combine a line-direction prefix (R for right, L for
left, and X for cross hatch) with a line-identification number:

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

See DISPLAY=
500 Chapter 6 • Plot Statements

FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
SOLID
each bar is filled with the color that is assigned to that bar.
GRADIENT
an alpha gradient is used to determine the bar fill color. Each bar is filled with a
color and a transparency gradient that starts at the bar top with the specified fill
color and transparency, and transitions to fully transparent at the bar baseline.
The initial fill color is determined by a style element or by the FILLATTRS=
option COLOR= suboption. The initial transparency is determined by the
DATATRANSPARENCY= option or by the FILLATTRS= option
TRANSPARENCY= suboption.

Interaction The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

Tip Use the DATATRANSPARENCY= option or the FILLATTRS=

option TRANSPARENCY= suboption to set the initial transparency
in the gradients.

See DATASKIN= on page 497

Default SOLID

Interaction The DISPLAY= option must include FILL for this option to have any
effect.

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that is
read.

Default All observations have a frequency count of 1.

Restriction If the value of the numeric-column is missing or is less than 1, then the
observation is not used in the analysis. If the value is not an integer,
then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation.

GROUP=column | discrete-attr-var | expression

creates a separate bar segment or bar for each unique group value in the specified
column.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.
HISTOGRAM Statement 501

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

For example, the Sashelp.Cars data contains a column named Origin, which
identifies the region that produces each car. This column could be used in the
HISTOGRAM statement to group the bars in the display:
proc template;
define statgraph histogram;
begingraph;
entrytitle "Highway Mileage Distribution by Origin";
layout overlay /;
histogram mpg_highway / name="histogram" group=origin;
discretelegend "histogram" / title="Origin:";
endlayout;
endgraph;
end;
run;

proc sgrender template=histogram data=sashelp.cars;

run;

Here is the output.

Defaults If bar fills or fill patterns are enabled by the ODS style or by the
DISPLAY= option, then each distinct group value is represented in the
plot by a different fill color or fill pattern. The fill colors are defined by
the Color attribute of the GraphData1–GraphDataN and GraphMissing
style elements. The fill patterns are defined by the FillPattern attribute of
the GraphData1–GraphDataN and GraphMissing style elements.

If bar outlines are enabled by the ODS style or by the DISPLAY= option,
then each distinct group value is represented in the plot by a different
outline. The outline colors are defined by the ContrastColor attribute of
the GraphData1–GraphDataN and GraphMissing style elements. The
outline thickness and pattern are defined by the LineThickness and
LineStyle attributes of the GraphOutlines style element
502 Chapter 6 • Plot Statements

Note The group values are mapped in the order in which they appear in the
data.

Tip You can individually override the representations that are used to identify
the groups. For example, in some ODS styles, each distinct group value is
represented by a different line pattern for the bar outlines. In that case,
you can use the PATTERN= setting on the OUTLINEATTRS= option to
assign the same line pattern to all of the bar outlines.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the MISSING= system option changes the default missing character, or
a user-defined format is applied to the group value. In those cases, the
attributes of the missing group value are determined by a GraphData1–
GraphDataN style element instead of by the GraphMissing style
element.

See “boolean ” on page 1339 for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.
HISTOGRAM Statement 503

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

NBINS=positive-integer
specifies the number of bins. This option can be used with the BINSTART= and
BINWIDTH= options to specify the bins.

Default Determined by the system.

Ranges A positive integer in SAS 9.4 and earlier releases

2–10000 starting with the first maintenance release of SAS 9.4

Note When BINWIDTH=, BINSTART=, or both are specified with this option
and the resulting data range is greater than the unbinned input data range,
the histogram might be padded with zero-height bins in order to
accommodate the BINSTART=, BINWIDTH=, and NBINS= option
values.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.

Default VERTICAL

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line properties of the bar outlines.

Defaults For non-grouped data, the GraphOutlines style element.

In the first maintenance release of SAS 9.4 and earlier releases, for
grouped data and filled bars, the ContrastColor and LineStyle attributes
of the GraphData1–GraphDataN style elements, and the LineThickness
attribute of the GraphOutlines style element.

Starting with the second maintenance release of SAS 9.4, for grouped
data and filled bars, the ContrastColor attribute of the GraphData1–
GraphDataN style elements, and the LineThickness and LineStyle
attributes of the GraphOutlines style element.

Interaction For this option to have any effect, the outlines must be enabled by the
ODS style or by the DISPLAY= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element.

“Line Options” on page 1349 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.
504 Chapter 6 • Plot Statements

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

SCALE=PERCENT | COUNT | PROPORTION | DENSITY

specifies the scale for the Y axis.

PERCENT displays the percentages between 0 and 100 on the Y axis.

COUNT displays the frequency counts on the Y axis.
PROPORTION displays the proportions between 0 and 1 on the Y axis.
DENSITY displays the true density estimates on the Y axis.

Default PERCENT

Tip When SCALE=PERCENT, the response axis tick values do not include a
percent sign after each value. To add a percent sign after each response axis
tick value in that case, use SCALE=PROPORTION, and then specify the
options shown in the following example for the response axis:
yaxisopts=(label="Percent" linearopts=(tickvalueformat=percent.))

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Note The X role represents the binned value. The Y role represents the
computed amount of X in the units specified by the SCALE= option.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Percent")
HISTOGRAM Statement 505

Default The column label or column name of the column assigned to the role.

Note The Y role represents the computed amount of X in the units specified by
the SCALE= option.

WEIGHT=numeric-column | expression
specifies a column that contains a bin-width calculation a priori weight for each
observation of the input data object.

Requirement The value must be nonnegative.

Interaction If the value for an observation is missing or is less than 1, then the
observation is removed from the analysis.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input X values represent midpoints, lower endpoints, or upper
endpoints of the bins. For example, if BINSTART=10 and BINWIDTH=10, then
using LEFTPOINTS would result in bins 10 - 20, 20 - 30, and so on. Using
RIGHTPOINTS would result in bins 0 - 10, 10 - 20, ...., and using MIDPOINTS
would result in bins 5 - 15, 15 - 25, ... .

Default MIDPOINTS

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
The histogram’s default bin width is computed by using the number of observations and
the range of the data. When a curve is overlaid on the histogram, the histogram’s bin
width is used to scale the curve so that the area under the curve is equal to the area of the
histogram.
The X-axis and Y-axis are linear by default. You can change axis properties with the
XAXISOPTS= and YAXISOPTS= options of the layout statement.
506 Chapter 6 • Plot Statements

Example: HISTOGRAM Statement

The following graph was generated by the “Example Program” on page 506:

Example Program
proc template;
define statgraph histogram;
begingraph;
entrytitle "Histogram of Vehicle Weights";
layout overlay /
xaxisopts=(label="Vehicle Weight (LBS)")
yaxisopts=(griddisplay=on);
histogram weight;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=histogram;

run;

HISTOGRAMPARM Statement
Creates a univariate histogram for specified values of bin midpoints and bin frequencies.
Restriction: Only uniform width bins are supported.
HISTOGRAMPARM Statement 507

Tip: Starting with the third maintenance release of SAS 9.4, you can use subpixel
rendering with this statement. It is enabled by default. To disable subpixel rendering,
specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 33. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.

Syntax
HISTOGRAMPARM X=numeric-column | expression
Y=non-negative-numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bars (outline and fill).
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display outlined bars, filled bars, or outlined and filled
bars.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the bars.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled bar area.
FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line properties of the bar outlines.
XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input X values represent midpoints, lower endpoints, or
upper endpoints of the bins.

Axes options
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

508 Chapter 6 • Plot Statements

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
histogram bin.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
specifies a column for the bar labels.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the bar labels.
DATALABELFITPOLICY=AUTO | NONE | ROTATE | SPLIT | SPLITALWAYS
specifies a policy for avoiding collisions among the bar labels when labels
are displayed.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin
endpoints (TRUE) or at the bin midpoints (FALSE).
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the column for the midpoint values.

Requirement The X values must be equally spaced.

Y=non-negative-numeric-column | expression
specifies the column for the frequency values. These values can be frequency counts,
percentages, or proportions between 0 and 1.

Requirement The Y values cannot be negative.

Optional Arguments
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.
HISTOGRAMPARM Statement 509

TRUE
specifies that the ENDLABELS= option determines how the axis ticks and value
labels are displayed.
FALSE
specifies that standard axes are used. Bin boundaries and midpoints that are set
by the ENDLABELS= option are ignored.

Default TRUE

Interactions This option is ignored when this histogram is not the primary plot. For
more information about primary plots, see “When Plots Share Data
and a Common Axis” on page 880.

BINAXIS=TRUE is ignored when the histogram is placed on a log

axis.

When this option is set to TRUE, some X-axis options that are set on
the parent layout might not apply, such as INTEGER=,
TICKVALUELIST=, TICKVALUESEQUENCE=, and
INCLUDERANGES=.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABEL=column | expression
specifies a column for the bar labels. The labels appear at the top or end of each bar,
depending on the chart orientation.

Default No data labels are displayed

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the bar labels.

Default The GraphDataText style element

Interactions The DATALABEL= option must be specified for this option to have
any effect.

When text options are specified, any font properties that are not
specified (color, family, size, weight, and style) are derived from the
GraphDataText style element.

See “General Syntax for Attribute Options” on page 1347

“Text Options” on page 1351

DATALABELFITPOLICY=AUTO | NONE | ROTATE | SPLIT | SPLITALWAYS

specifies a policy for avoiding collisions among the bar labels when labels are
displayed.
AUTO
selects a collision avoidance policy based on the chart orientation and data type.
For a numeric column with ORIENT=VERTICAL, AUTO rotates the labels if
they do not fit the midpoint spacing. For a character column, AUTO splits the
labels if they do not fit the midpoint spacing.

Note When ORIENT=HORIZONTAL, AUTO always draw the labels

horizontally.
510 Chapter 6 • Plot Statements

Tip If character labels do not fit after splitting, then try using ROTATE instead
of AUTO.

See ORIENT= on page 516 for information about chart orientation.

NONE
does not attempt to fit bar labels that collide.
ROTATE
rotates the bar labels for vertical bars if the labels collide in the available width.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

SPLIT
splits the label for vertical bars at a split character only if a split is needed at that
character in order to make the label fit the available space. No split occurs at split
characters that occur where a split is not needed. If the label does not contain any
of the specified split characters, then a split does not occur. In that case, if the
label does not fit the available space, then it might collide with the adjoining
labels.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

See DATALABELSPLITCHAR= for information about specifying

the split characters

SPLITALWAYS
splits the label for vertical bars at every occurrence of a split character. If the
label does not contain any of the specified split characters, then a split does not
occur.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

See DATALABELSPLITCHAR= for information about specifying

the split characters

Default AUTO

Requirement The DATALABEL= option must also be specified.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split. When multiple
split characters are specified, each character in the list is treated as a separate split
character unless the specified characters appear consecutively in the data label. In
that case, all of the specified split characters together are treated as a single split
character.
When DATALABELFITPOLICY=SPLIT, data labels are split on a split character
only if a split is needed at that point in order to make the label fit. When
DATALABELFITPOLICY=SPLITALWAYS, the data labels are split
unconditionally on each occurrence of a split character. If the data label does not
contain any of the specified characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

HISTOGRAMPARM Statement 511

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

The DATALABELFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interactions The DATALABELFITPOLICY= option specifies the policy that is

used to manage the split behavior of the data labels.

The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
TRUE
drops a split character from the data label when a split occurs at that character.
Split characters at which a split does not occur are left in place. The
DATALABELFITPOLICY= option determines where the labels are split. When
DATALABELFITPOLICY=SPLIT, each label is split at a split character only
where a split is needed in order to make the label fit the available space. At each
split point, the split character is dropped, and the characters that follow the split
character, up to but not including the split character at the next split point, are
wrapped to the following line.
When DATALABELFITPOLICY=SPLITALWAYS, each label is split at every
instance of a split character. All of the split characters are dropped. The
characters that follow each split character, up to but not including the next split
character, are wrapped to the next line.
The following figure shows how label “Product*Group*1” is split when the
DATALABELSPLITCHARDROP=TRUE and DATALABELSPLITCHAR="*"
options are used with the SPLIT and SPLITALWAYS fit policies.

In this example, when DATALABELFITPOLICY=SPLIT, the label is split at the

first occurrence of the asterisk in order to make the label fit. No split is needed at
the second asterisk. The first asterisk is dropped, and Group*1 wraps to the next
line. Notice that the second asterisk is not dropped in this case. When
DATALABELFITPOLICY=SPLITALWAYS, the label is split at every
occurrence of the asterisk. In this case, both asterisks are dropped, and the
characters that follow each asterisk wrap to the next line.
512 Chapter 6 • Plot Statements

FALSE
includes the split characters in the data label. The DATALABELFITPOLICY=
option determines how the split characters are displayed. When
DATALABELFITPOLICY=SPLIT, each data label is split at a split character
only where a split is needed in order to make the label fit the available space. A
split might not occur at every split character in the label. At each split point, the
split character remains as the last character in the current line. The characters that
follow the split character, up to and including the split character at the next split
point, are then wrapped to the following line. This process repeats until the entire
data label is displayed.
When DATALABELFITPOLICY=SPLITALWAYS, each data label is split at
every instance of a split character in the label regardless of whether a split is
actually needed. Each split character remains as the last character in the current
line. The characters that follow each split character, up to and including the next
split character, are then wrapped to the next line.
The following figure shows how label “Product*Group*1” is split when the
DATALABELSPLITCHARDROP=FALSE and DATALABELSPLITCHAR="*"
options are used with the SPLIT and SPLITALWAYS fit policies.

In this example, when DATALABELFITPOLICY=SPLIT, the label is split at the

first occurrence of the asterisk in order to make the label fit. No split is needed at
the second asterisk. The characters that follow the first asterisk wrap to the next
line. When DATALABELFITPOLICY=SPLITALWAYS, the label is split at
every occurrence of the asterisk. Each asterisk remains as the last character in the
current line, and the characters that follow are wrapped to the next line.

Default TRUE. A split character is dropped from thedata label when a split
occurs at that character.

Requirements The DATALABEL= option must also be specified.

The DATALABELFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows
histogram bars with each of the skins applied.
HISTOGRAMPARM Statement 513

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

When FILLTYPE=GRADIENT is in effect, DATASKIN=SHEEN is

ignored. In that case, use one of the other skins.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bars (outline and fill).

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the bar
fills. You can combine this option with FILLATTRS= to set one
transparency for the bar outlines but a different transparency for the bar
fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display outlined bars, filled bars, or outlined and filled bars.
514 Chapter 6 • Plot Statements

STANDARD
displays outlined, filled bars. (Currently, the same as ALL.)
ALL
displays outlined, filled bars.
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:
OUTLINE
displays outlined bars.
FILL
displays bars with a color fill.
FILLPATTERN
displays bars with a patterned fill.

Note This option is valid in the first maintenance release of SAS 9.4 and
later releases.

Default The GraphHistogram:DisplayOpts style reference.

Tip Use the OUTLINEATTRS= , FILLATTRS= , and FILLPATTERNATTRS=

options to control the appearance of the bars.

ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin endpoints
(TRUE) or at the bin midpoints (FALSE).

Default FALSE.

Interactions This option is ignored if this plot is not the primary plot in the parent
layout. For more information, see the PRIMARY= option.

This option is ignored if BINAXIS= FALSE. By default,

BINAXIS=TRUE.

If the TICKS= suboption is specified in the XAXISOPTS= option,

then this option is ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the bars.

Default The GraphDataDefault style element

Interaction For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

Tip The DATATRANSPARENCY= option sets the transparency for bar fills
and bar outlines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines but
a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
HISTOGRAMPARM Statement 515

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled bar area.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
style-element
specifies the name of a style element. You can specify only one of the elements
GraphData1–GraphDataN.

Restriction The only styles that are delivered by SAS that support fill patterns
are JOURNAL2, JOURNAL3, and MONOCHROMEPRINTER. If
any other such style is in effect and this option uses style-element in
its specification, then this option is ignored.

(fill-pattern-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:
COLOR=color | style-reference
specifies a color to use for the bar-fill-pattern lines.
PATTERN=line-pattern
specifies a line pattern to use for the bar fill.
To specify a line-pattern, combine a line-direction prefix (R for right, L for
left, and X for cross hatch) with a line-identification number:

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

See DISPLAY=

FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
SOLID
each bar is filled with the color that is assigned to that bar.
516 Chapter 6 • Plot Statements

GRADIENT
an alpha gradient is used to determine the bar fill color. Each bar is filled with a
color and a transparency gradient that starts at the bar top with the specified fill
color and transparency, and transitions to fully transparent at the bar baseline.
The initial fill color is determined by a style element or by the FILLATTRS=
option COLOR= suboption. The initial transparency is determined by the
DATATRANSPARENCY= option or by the FILLATTRS= option
TRANSPARENCY= suboption.

Interaction The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

Tip Use the DATATRANSPARENCY= option or the FILLATTRS=

option TRANSPARENCY= suboption to set the initial transparency
in the gradients.

See DATASKIN= on page 512

Default SOLID

Interaction The DISPLAY= option must include FILL for this option to have any
effect.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.

Default VERTICAL

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line properties of the bar outlines.

Default The GraphOutlines style element.

Interaction For this option to have any effect, the outlines must be enabled by the
ODS style or the DISPLAY= option.
HISTOGRAMPARM Statement 517

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, and DATALABEL.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a histogram
bin. If this option is used, then it replaces all of the information that is displayed by
default. Roles for columns that do not contribute to the histogram can be specified
along with roles that do.
(role-list)
an ordered, space-separated list of unique HISTOGRAMPARM and user-defined
roles. HISTOGRAMPARM roles include X, Y, and DATALABEL.

Tip User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns assigned to
the roles X and Y, as well as the column Pct, which is not assigned to
any pre-defined HISTOGRAMPARM role. The Pct column must first
be assigned a role:
518 Chapter 6 • Plot Statements

ROLENAME=(TIP1=PCT)
TIP=(X Y TIP1)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X and Y.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.
Example: HISTOGRAMPARM Statement 519

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input X values represent midpoints, lower endpoints, or upper
endpoints of the bins.

Default MIDPOINTS

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
When a curve is overlaid on the histogram, the histogram bin width is used to scale the
curve so that the area under the curve is equal to the area of the histogram.
The X-axis and Y-axis are linear by default. You can change axis properties with the
XAXISOPTS= and YAXISOPTS= options of the LAYOUT OVERLAY statement.

Example: HISTOGRAMPARM Statement

The following graph was generated by the “Example Program” on page 520:
520 Chapter 6 • Plot Statements

Example Program
proc template;
define statgraph histogramparm;
begingraph;
entrytitle "Histogram of Vehicle Weights";
layout overlay;
histogramparm x=midpoint y=frequency;
endlayout;
endgraph;
end;
run;

data bins;
input Midpoint Frequency;
label midpoint="Vehicle Weight (LBS)";
datalines;
2000 18
3000 208
4000 159
5000 36
6000 6
7000 1
;
run;

proc sgrender data=bins template=histogramparm;

run;
LINECHART Statement 521

LINECHART Statement
Creates a line chart that is computed from input data.
Restrictions: GROUPDISPLAY= STACK has stringent requirements for the data. If the
requirements are not met for all the data, then the line chart is not drawn. If the
LINECHART statement is in a LAYOUT DATALATTICE or LAYOUT DATAPANEL
block, then the requirements are tested for all of the data, not for each individual
panel.
The category axis (the X-axis when ORIENT=VERTICAL or the Y-axis when
ORIENT=HORIZONTAL) is always discrete.
The response axis (the Y-axis when ORIENT=VERTICAL or the X-axis when
ORIENT=HORIZONTAL) is always numeric.
The LINECHART statement performs discrete binning for a numeric category column
only.
Note: Specifying only the CATEGORY= option creates a computed line chart with vertices
representing frequency counts or percents of unique CATEGORY values. For a non-
grouped chart, specifying both the CATEGORY= and RESPONSE= options creates
a computed line chart with vertices representing the summarized values of the
RESPONSE values that are categorized by unique CATEGORY values.
Tips: The line segments in the chart always join the categorical values in the order in
which they appear on the axis. By default, the vertices in the line chart appear in the
order in which the X values occur in the input data. To change the categorical axis
tick value order, use the SORTORDER= or TICKVALUELIST= suboption of the
DISCRETEOPTS= option for the X-axis. For example:
XAXISOPTS=(DISCRETEOPTS=(SORTORDER=ASCENDINGFORMATTED))

The response axis of the line chart includes the zero value by default. You can use
the BASELINEINTERCEPT= to change the Y intercept.
By default, missing category values are ignored. To treat missing category values as
a category, include the INCLUDEMISSINGDISCRETE=TRUE option in the
BEGINGRAPH statement.
Starting with the third maintenance release of SAS 9.4, subpixel rendering is enabled
by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page
33.For information about the ODS GRAPHICS statement SUBPIXEL= option, see
“ODS GRAPHICS Statement” in SAS ODS Graphics: Procedures Guide.
See: “LAYOUT DATALATTICE Statement” on page 45
“LAYOUT DATAPANEL Statement” on page 70
Chapter 8, “Axis Options in Layouts,” on page 889
“BEGINGRAPH Statement” on page 21

Syntax
LINECHART CATEGORY=column | expression </option(s)> ;
LINECHART CATEGORY=column | expression
RESPONSE=numeric-column | expression </option(s) > ;
522 Chapter 6 • Plot Statements

Summary of Optional Arguments

Appearance options
BREAK=TRUE | FALSE
breaks the line at missing values of the RESPONSE variable.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the line, marker,
and fill colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the lines and markers.
DATATRANSPARENCY=number
specifies the degree of the transparency of the line, markers, and vertex
labels, if displayed.
DISPLAY=STANDARD | ALL | (display-options)
specifies which graphical features to display.
FILLATTRS=style-element | (fill-options)
specifies the appearance of the filled area.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
GROUPDISPLAY=OVERLAY | STACK
specifies how to display grouped lines.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line
pattern) and fill attributes to one of the GraphData1–GraphDataN style
elements.
LINEATTRS=style-element | (line-options)
specifies the appearance of the line.
MARKERATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the data markers.
MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.
MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis.
SMOOTHCONNECT=TRUE | FALSE
specifies that a smoothed line passes through all vertices.

Axes options
BASELINEINTERCEPT=number | AUTO | AXISMIN | AXISMINEXTEND |
AXISMAX | AXISMAXEXTEND
specifies the Y-intercept for the baseline.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
LINECHART Statement 523

specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
VERTEXLABEL=TRUE | FALSE
specifies whether to label the vertices with their response value (or statistic).
VERTEXLABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the vertex labels.
VERTEXLABELFORMAT=format
specifies the format used to display the vertex label response or statistic.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate line for each unique group value in the specified column.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a line segment, marker, or fill area
is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
STAT=FREQ | PCT | SUM | MEAN | PROPORTION
specifies the statistic to be computed for the Y-axis.

Required Argument
CATEGORY=column | expression
specifies the column or expression for the category values. Duplicated category
values are summarized into a unique value. All values are treated as discrete.
524 Chapter 6 • Plot Statements

Optional Arguments
BASELINEINTERCEPT=number | AUTO | AXISMIN | AXISMINEXTEND |
AXISMAX | AXISMAXEXTEND
specifies the Y-intercept for the baseline.
number
specifies the Y coordinate of the baseline. This value is included in the data range
that is reported by the line chart.

Interaction When number is specified, if necessary, the response axis data

range is extended to include the baseline intercept. When a
logarithmic response axis is requested and number is 0 or a negative
value, the response axis reverts to a linear axis. To restore the log
axis in that case, set BASELINEINTERCEPT= to a positive value.

AUTO
bases the intercept on the response axis range in the following ways:
• If the response axis range has both positive and negative values or contains 0,
then the intercept is 0.
• If the response axis range contains all positive values, then AUTO is
interpreted as AXISMINEXTEND.
• If the response axis range contains only negative values, then AUTO is
interpreted as AXISMAXEXTEND.
AXISMIN
places the baseline at the minimum value of the axis range.
AXISMINEXTEND
places the baseline at the start of the minimum offset in the wall. This location
corresponds to the bottom edge of the wall when there is no inner margin plot
and the axis is not reversed. If there is an inner margin plot at the bottom, then
the baseline is placed at the boundary of the inner margin and the minimum
offset.
AXISMAX
places the baseline at the maximum value of the axis range.
AXISMAXEXTEND
places the baseline at the start of the maximum offset in the wall. This location
corresponds to the top edge of the wall when there is no inner margin plot, and
the axis is not reversed. If there is an inner margin plot at the top, then the
baseline is placed at the boundary of the inner margin and the maximum offset.

Default AUTO

Interactions When GROUPDISPLAY= STACK is in effect, this option is ignored,

and the plot is drawn as if BASELINEINTERCEPT=0.

When DISPLAY= includes FILL, the fill extends to the baseline that
is specified by the BASELINEINTERCEPT= option.

BREAK=TRUE | FALSE
breaks the line at missing values of the RESPONSE variable.

Default FALSE
LINECHART Statement 525

Note If BREAK=FALSE, then missing values are skipped and a continuous line
is drawn.

See “boolean ” on page 1339 for other Boolean values that you can use.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the line, marker, and
fill colors.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Requirement The COLORRESPONSE values should remain constant for each

group value in a grouped plot and for the entire plot in an ungrouped
plot. If the COLORRESPONSE column has multiple values for a
526 Chapter 6 • Plot Statements

single GROUP value or for a non-grouped plot, unexpected results

might occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

When this option is specified without the GROUP= option, only a

single line is generated for the plot, and the line color is derived from
the COLORRESPONSE= value.

When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map. The line and marker colors in that
case are controlled by the contrastColor attribute of the
GraphDataDefault style element.

When fill is not displayed, this option overrides suboption COLOR=

in the LINEATTRS= and MARKERATTRS= options, and varies the
line and marker colors according to the color gradient or the attribute
map.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

For a numeric column or expression, the ThreeColorRamp style

element defines the fill color gradient, and the ThreeColorAltRamp
style element defines the line color gradient.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the lines and markers. The following figure shows
lines and CIRCLEFILLED markers with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
LINECHART Statement 527

specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The skin appearance is based on the color that is in effect for the lines
and markers.

DATATRANSPARENCY=number
specifies the degree of the transparency of the line, markers, and vertex labels, if
displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the fills.
You can combine this option with FILLATTRS= to set one transparency for
the lines but a different transparency for the fills. For example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies which graphical features to display.
STANDARD
displays a line with no markers and no fill under the line to the baseline.
ALL
displays a line with markers and the fill under the line to the baseline.
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:
FILL
displays a filled area between the line and the baseline.

Interaction When GROUPDISPLAY=STACK, DISPLAY=FILL fills

between adjacent group lines except for the first group, which
fills to the baseline.

LINE
displays line segments that join the vertices.
MARKERS
displays markers at each vertex.

Default STANDARD

Tip Use the LINEATTRS=, MARKERATTRS=, and FILLATTRS= options to

control the appearance of the line segments, markers, and fill, respectively.

FILLATTRS=style-element | (fill-options)
specifies the appearance of the filled area.

Defaults For non-grouped data, the COLOR attribute of GraphDataDefault

style element.
528 Chapter 6 • Plot Statements

For grouped data, the COLOR attribute of GraphData1–GraphDataN

style elements is used.

Interactions For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is
ignored, and the fill colors vary according to the gradient.

Note When style-element is specified, only the style element’s COLOR

attribute is used.

Tip The DATATRANSPARENCY= option sets the transparency for the

fills and the lines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the lines but a
different transparency for the fills. For example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347

“Fill Options” on page 1348

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline. When this option is TRUE, the fill color and outline color for
filled markers are determined in the following ways:
• If the GROUP= option is specified, then by default, the fill color is derived
from the GraphData1–GraphDataN style elements Color attribute, and the
marker outlined color is derived from the GraphData1–GraphDataN style
elements ContrastColor attribute.
• If the GROUP= option is not specified, then by default, the fill color is
derived from the GraphDataDefault style elements Color attribute, and the
marker outlined color is derived from the GraphOutline style elements
ContrastColor attribute.
FALSE
draws the markers using fill or an outline, but not both.

Default FALSE

Tip To specify the marker fill and outline colors for a non-grouped plot, set this
option to TRUE, and then use the MARKERFILLATTRS= and
MARKEROUTLINEATTRS= options to specify the colors.

See GROUP= on page 529

MARKERFILLATTRS= on page 532

MARKEROUTLINEATTRS= on page 533

“boolean ” on page 1339 for other Boolean values that you can use.
LINECHART Statement 529

GROUP=column | discrete-attr-var | expression

creates a separate line for each unique group value in the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value is represented in the plot by a different

combination of color, line pattern, and marker symbol. Lines and
markers vary according to the ContrastColor, LineStyle, and
MarkerSymbol attributes of the GraphData1–GraphDataN and
GraphMissing style elements.

Interaction When both the GROUP= and COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.

Tip The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the lines, but you can use the
PATTERN= suboption of the LINEATTRS= option to assign the same
line pattern to all lines.

See “DISCRETEATTRMAP Statement” on page 1287

GROUPDISPLAY=OVERLAY | STACK
specifies how to display grouped lines.
OVERLAY
displays group values overlaid on top of each other.
STACK
displays group values as stacked lines.

Default OVERLAY

Restriction When STACK is in effect, if any response value is negative or if any

crossing of the group value with the category is absent or is a missing
value, then the chart is not drawn and a warning message is written to
the SAS log.

Interaction When STACK is in effect, the BASELINEINTERCEPT= option is

treated as if it is set to zero.

Tip When the response axis is linear, STAT=MEAN or STAT=PCT, and

GROUPDISPLAY=STACK, the axis tick values might be displayed as
integer values. When the response axis is linear, STAT=MEAN or
STAT=PCT, and GROUPDISPLAY=OVERLAY, the axis tick values
might be displayed as decimal values. To keep the integer axis values
for both cases, you can specify the INTEGER=TRUE option for the
response axis. See INTEGER= on page 915.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
530 Chapter 6 • Plot Statements

DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
DESCENDING
orders the groups within a category in descending order.

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line pattern)
and fill attributes to one of the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.
LINECHART Statement 531

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default If the RESPONSE= option is specified, then the response variable label
is used. Otherwise, the CATEGORY= variable label is used. If a label
is not assigned to the response variable or category variable, then the
variable name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | (line-options)
specifies the appearance of the line.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the LineThickness attributes of the GraphDataDefault

style element, and the ContrastColor and LineStyle attributes of the
GraphData1–GraphDataN style elements.

Interaction When COLORRESPONSE= is in effect, the LINEATTRS= suboption

COLOR= is ignored, and the line fill colors vary according to the
gradient.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

532 Chapter 6 • Plot Statements

MARKERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes of

the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions If FILLEDOUTLINEDMARKERS=TRUE, then this option’s

COLOR= suboption is ignored. In that case, to specify the marker fill
color, use the MARKERFILLATTRS= option instead.

This option’s COLOR= suboption overrides the default behavior for

grouped data. When the COLOR= suboption is specified in that case,
all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior for

grouped data. When the SYMBOL= suboption is specified in that
case, all markers have the same symbol, and the symbol color alone
distinguishes the markers.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

This option is ignored if the DISPLAY= option disables the display of

the markers.

If the DATASKIN= option is in effect, then the data skin determines

the marker outlines. Any outline-related settings from the current ODS
style or from the marker attribute options are ignored.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Marker Options” on page 1350 for available marker-options.

MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.

Defaults For non-grouped data, the COLOR attribute of the GraphDataDefault

style element

For grouped data, the COLOR attribute of a GraphData1–GraphDataN

style element

Restriction The TRANSPARENCY= fill option is ignored. Use the

MARKERATTRS= option to set the marker transparency.

Interaction This option is in effect only when

FILLEDOUTLINEDMARKERS=TRUE and the DISPLAY= option
enables fill display.
LINECHART Statement 533

Note When style-element is specified, only the style element’s COLOR

attribute is used.

See “General Syntax for Attribute Options” on page 1347

“Fill Options” on page 1348

MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, the LineThickness attritube of the GraphOutlines

style element and the ContrastColor attribute of a GraphData1–
GraphDataN style element.

Restriction The line style of the marker outline is always solid.

Interaction This option is ignored when a data skin is applied by the current style
or by the DATASKIN= option. In the latter case, the outline is set by
the data skin.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR and LINETHICKNESS attributes are used.

See “General Syntax for Attribute Options” on page 1347

“Line Options” on page 1349

NAME="string"
assigns a name to this plot statement for reference in other template statements. This
option is used mostly in the CONTINUOUSLEGEND on page 1098 statement in
order to coordinate the use of colors and line patterns between the plot and the
legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis.

Default VERTICAL

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
534 Chapter 6 • Plot Statements

the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles CATEGORY, RESPONSE, DATALABEL, and
GROUP.

SMOOTHCONNECT=TRUE | FALSE
specifies that a smoothed line passes through all vertices.

Default FALSE. Straight line segments are used if the vertices are to be
connected.

Interaction This option is ignored when GROUPDISPLAY=STACK.

See “boolean ” on page 1339 for other Boolean values that you can use.

STAT=FREQ | PCT | SUM | MEAN | PROPORTION

specifies the statistic to be computed for the Y-axis. For bar charts with no
RESPONSE= column:

FREQ frequency count

PCT percentages between 0 and 100 inclusive
PROPORTION proportions between 0 and 1 inclusive
For bar charts with a RESPONSE= column:

SUM
MEAN

Defaults SUM for line charts that specify the RESPONSE= argument.

FREQ for line charts that do not specify the RESPONSE= argument.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a line. If this
option is used, then the information specified replaces all of the information that is
LINECHART Statement 535

displayed by default. You can specify roles for columns that do not contribute to the
line chart along with roles that do.
(role-list)
an ordered, space-separated list of unique LINECHART roles and user-defined
roles. LINECHART roles include CATEGORY, RESPONSE,
COLORRESPONSE, DATALABEL, and GROUP.

Note The COLORRESPONSE role is valid starting with the third maintenance
release of SAS 9.4.

Tip User-defined roles are defined with the ROLENAME= option.

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to the following roles are automatically

included in the data tip information: CATEGORY, RESPONSE, and
GROUP.

Restriction Data tips are available only for graphs that are written to the ODS
HTML destination.

Requirement To generate data tips, include an ODS GRAPHICS ON statement that

specifies the IMAGEMAP option. See “ODS GRAPHICS Statement”
in SAS ODS Graphics: Procedures Guide for information about the
ODS GRAPHICS statement.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Note The RESPONSE role represents the computed values for the
CATEGORY role (and RESPONSE= role), based on the STAT=
option.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

Example To display data tips for the columns assigned to the roles X and Y as
well as the user-defined role TIP1:
ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.
536 Chapter 6 • Plot Statements

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a line segment, marker, or fill area is
selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
line that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable line segments, markers, and fill areas, you
must include an ODS GRAPHICS ON statement that specifies the
IMAGEMAP option, and you must write the output to the ODS
HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Notes For non-grouped data, the values of the column are expected to be
same for each unique X value. If they are not, then the results might
be unpredictable.

For grouped data, the values of the column are expected to be the
same for each unique X and GROUP combination.

Tips The URL value can be blank for some X values, meaning that no
action is taken when the line segments for those X values are
selected.
LINECHART Statement 537

The URL value can be the same for different X values, meaning that
the same action is taken when the line segments for those X values
are selected.

VERTEXLABEL=TRUE | FALSE
specifies whether to label the vertices with their response value (or statistic).

Default FALSE

Interaction When GROUPDISPLAY=STACK is in effect, vertex labeling displays

the sum of the vertex responses per category.

Tips You can modify the visual attributes for the label by using the
VERTEXLABELATTRS= option.

You can modify the text format by using the

VERTEXTLABELFORMAT= option.

See “boolean ” on page 1339 for other Boolean values that you can use.

VERTEXLABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the vertex labels.

Default The GraphDataText style element.

Requirement VERTEXLABEL=TRUE must be in effect for this option to have any

effect.

Interaction If one or more text options are specified and they do not include all
the font properties (color, family, size, weight, style), then non-
specified properties are derived from the GraphDataText style
element.

Note When style-element is specified, only the style element’s COLOR,

FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT
attributes are used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

VERTEXLABELFORMAT=format
specifies the format used to display the vertex label response or statistic.

Default The column format assigned to the response variable, or BEST6.2 if

the column does not have an associated format.

Requirement VERTEXLABEL=TRUE must be in effect for this option to have any

effect.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X
538 Chapter 6 • Plot Statements

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Optional Response Argument

RESPONSE=numeric-column | expression
specifies the numeric column or expression for the response values.

Details
A line chart shows the relationship of one variable to another as trends in the data over a
period of time. The trends are shown by connecting the successive data points with a
line. Typically, a line chart is used to chart a response value against a discrete categorical
value where each value on the horizontal axis has only one corresponding value on the
vertical axis. A grouping variable can be used to show multiple trends based on the
group values.
In a line chart, the category axis (X) is always discrete, and the response axis (Y) is
always linear. The line segments in the chart always join the categorical values in the
order in which they appear on the axis. The vertices in the line chart appear in the order
in which the categorical values appear in the input data.

Examples

Example 1: Grouped Line Chart with Custom Line and Fill Attributes
This example shows you how to generate a simple line chart by using the LINECHART
statement and how to use the statement options to customize the plot.
Example 1: Grouped Line Chart with Custom Line and Fill Attributes 539

The following graph was generated by the “Example Program” on page 539. It shows
the trend of the average closing price of the IBM, Intel, and Microsoft stocks from 1995
through 2005.

Example Program
Here is the SAS code for this example.
/* Extract per-year data for 1995 through 2005 from SASHELP.STOCKS */
data linechartdata;
set sashelp.stocks(where=(year(date) between 1995 and 2005));
year=year(date);
label year="Year";
run;

/* Define the line chart template */

proc template;
define statgraph linechart;
begingraph;
entrytitle "Stock Index Performance: 1995 - 2005";
layout overlay /
/* Add a grid */
xaxisopts=(griddisplay=on gridattrs=(pattern=dot
color=lightgray))
yaxisopts=(griddisplay=on gridattrs=(pattern=dot
color=lightgray));

/* Generate the line chart */

linechart category=year response=close / name="linechart"

/* Compute the mean statistic */

stat=mean

/* Group by stock to draw a line for each stock */

540 Chapter 6 • Plot Statements

group=stock

/* Set the baseline at 0 */

baselineintercept=0

/* Display the lines and fill */

display=(line fill)

/* Specify the line attributes */

lineattrs=(thickness=2)

/* Specify the fill attributes */

fillattrs=(transparency=0.8);

/* Add a legend */
discretelegend "linechart";
endlayout;
endgraph;
end;
run;

/* Generate the chart */

proc sgrender data=linechartdata template=linechart;
run;

Example 2: Grouped Line Chart with Discrete Attribute Map

This example shows you how to create a more flexible template that enables you to
easily generate multiple line graphs that use different data. It also shows you how to
customize the plot appearance by using a discrete attribute map.
The following graph was generated by the “Example Program” on page 541. It shows
the trend of the average monthly closing price of the IBM, Intel, and Microsoft stocks
for 2001.
Example 2: Grouped Line Chart with Discrete Attribute Map 541

Example Program
Here is the SAS code for this example.
/* Define the line chart template */
proc template;
define statgraph linechart;
begingraph;
/* Create a dynamic variable for the year */
dynamic year;

/* Define the display attributes for each stock. Since

DISCRETEATTRMAP does not support the MARKERATTRS=
suboption SIZE=, the marker size is set separately. */
discreteattrmap name="stocks" / ignorecase=true ;
value "IBM" /
markerattrs=(color=blue symbol=trianglefilled)
lineattrs=(color=lightblue pattern=solid);
value "Intel" /
markerattrs=(color=red symbol=circlefilled)
lineattrs=(color=verylightred pattern=solid);
value "Microsoft" /
markerattrs=(color=orange symbol=squarefilled)
lineattrs=(color=verylightorange pattern=solid);
enddiscreteattrmap ;

/* Associate the attribute map with input data column Stock and
assign the name STOCKATTRS to the named association */
discreteattrvar attrvar=stockattrs var=stock attrmap="stocks" ;

entrytitle "Stock Index Performance in " year;

layout overlay /
/* Add a grid */
xaxisopts=(griddisplay=on gridattrs=(pattern=dot
color=lightgray))
yaxisopts=(griddisplay=on gridattrs=(pattern=dot
color=lightgray));
linechart category=date response=close / name="linechart"
/* Compute the mean */
stat=mean

/* Group by stock using the specified attributes */

group=stockattrs

/* Display the lines and markers */

display=(line markers)

/* Set the marker size */

markerattrs=(size=6)

/* Show vertex labels and specify their attributes */

vertexlabel=true
vertexlabelattrs=(size=7pt)
vertexlabelformat=dollar4.0;

/* Add a legend */
discretelegend "linechart";
542 Chapter 6 • Plot Statements

endlayout;
endgraph;
end;
run;

/* Create a macro that generates a line chart for a specific year */

%macro genchart(year=);
/* Generate the chart */
proc sgrender data=sashelp.stocks template=linechart;
where year(date)=&year; /* Get the data for the specified year */
format date monname3.; /* Format the date as 3-letter month */
dynamic year="&year;"; /* Pass the year to the template */
run;
%mend genchart;

/* Generate a chart for 2001 */

%genchart(year=2001);

LINEPARM Statement
Creates a straight line specified by a point and a slope.
Requirement: A LINEPARM statement can be used only within a 2-D layout (OVERLAY,
OVERLAYEQUATED, DATALATTICE, or DATAPANEL). Another plot statement that
is derived from data values that provide boundaries for the axis area must be
included.

Syntax
LINEPARM X=number | numeric-column | expression
Y=number | numeric-column | expression
SLOPE=number | numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the data for the line is considered when the data ranges are
determined for the axes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the line.
EXTEND=TRUE | FALSE
specifies whether the line is to be drawn to the area bounded by the axes.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the line.

Axes options
XAXIS=X | X2
LINEPARM Statement 543

specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
CURVELABEL="string" | column | expression
specifies a label for the line.
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the line label.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN
specifies the position of the line label relative to the line end points.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the line label at the specified split characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the line label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the line label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the line label block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate parameterized line plot for each unique group value of the
specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=number | numeric-column | expression
specifies the X coordinate of a point.
By default, if the specified value is outside of the data range, then the data range is
extended to include the specified intercept. This behavior can be changed with the
CLIP= option. If a numeric-column is specified and the column contains missing
values, then no line is drawn for the missing values.
Values are displayed in the units of the data.
Y=number | numeric-column | expression
specifies the Y coordinate of a point.
544 Chapter 6 • Plot Statements

By default, if the specified value is outside of the data range, then the data range is
extended to include the specified intercept. This behavior can be changed with the
CLIP= option. If a numeric-column is specified and the column contains missing
values, then no line is drawn for the missing values.
Values are displayed in the units of the data.
SLOPE=number | numeric-column | expression
specifies the slope of the line. Slope can be positive or negative.
SLOPE=0 creates a line parallel to the X-axis. SLOPE=. (a missing value) creates a
line parallel to the Y-axis.

Optional Arguments
CLIP=TRUE | FALSE
specifies whether the data for the line is considered when the data ranges are
determined for the axes.
FALSE
specifies that the data for the line contributes to the data range for each axis.
Each axis might be extended in order to force the display of the line. When
CLIP=FALSE, the SLOPE= option determines how the X= and Y= values
contribute to the axis data range in the following ways:
• If SLOPE=0, then only the Y= values contribute to the axis data range.
• If SLOPE=. (missing), then only the X= values contribute to the axis data
range.
• If SLOPE= is neither 0 nor missing, then the X= and Y= values contribute to
the axis data range.
TRUE
specifies that the data for the line is ignored when axis scales are being
established. Each axis scale is determined by the other plots in the layout. In this
case, the line might not be displayed if its data range is not within the data ranges
of the other plots.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABEL="string" | column | expression

specifies a label for the line.

Default No line label is displayed

Restrictions When the GROUP= option is specified, "string" and expression are
not valid. Use column in that case.

When the GROUP= option is not specified, column is not valid. Use
"string" or expression in that case.

The line label for missing values is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

See GROUP= on page 549

LINEPARM Statement 545

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the line label.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, text color is derived from the

GraphData1:ContrastColor–GraphDataN:ContrastColor style
references. The font is derived from the GraphValueText style
element.

Interactions For this option to take effect, the CURVELABEL= option must also
be used.

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

Tip When the GROUP= option is used, each distinct group value might be
represented by a different color. The line label that is associated with
the group is assigned the group color. This option can be used to
specify a single color for all line labels in a plot, without affecting the
line colors.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the line label relative to the plot area.
INSIDE
locates the labels inside the plot area
OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the LINEPARM is used in multi-cell

layouts such as LATTICE, DATAPANEL, or DATALATTICE, where
axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the line labels
appear.

See “Location and Position of Curve Labels” on page 185

CURVELABELPOSITION=AUTO | MAX | MIN

specifies the position of the line label relative to the line end points. This option is
used in conjunction with the CURVELABELLOCATION= option to determine
where the line label appears.
546 Chapter 6 • Plot Statements

AUTO
automatically positions the line label near the line boundary along unused axes
whenever possible (typically Y2 and X2) in order to avoid collision with tick
values.

Restriction This option is used only when CURVELABELLOCATION=

OUTSIDE.

MAX
forces the line label to appear near maximum line values (typically, upper right).
MIN
forces the line label to appear near minimum line values (typically, lower left).

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

MAX when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION= INSIDE

is specified.

Interaction For this option to take effect, the CURVELABEL= option must also be
specified.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN=

in an axis statement, the data points that are used to determine the
position of the curve label might fall outside of the graph area. In that
case, the curve label might not be displayed or might be positioned
incorrectly.

See “Location and Position of Curve Labels” on page 185

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the line label at the specified split characters. When a line
label is split, the label is split on each occurrence of the specified split characters.

Default FALSE. The line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the line label can be split if needed. When
multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
line label. In that case, all of the specified split characters together are treated as a
single split character.
When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the line
label is split unconditionally at each occurrence of any of the specified split
characters. If the line label does not contain any of the specified characters, then the
label is not split.
LINEPARM Statement 547

"character-list"
one or more characters with no delimiter between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the following
option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the line label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the line label text.
TRUE
drops the split characters from the line label text.
FALSE
includes the split characters in the line label text. When
CURVELABELSPLIT=TRUE and
CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split character, up
to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a line label with the following
specifications:
• CURVELABELPOSITION=MAX
• CURVELABEL="Product*Group*A"
• CURVELABELSPLIT=TRUE
• CURVELABELSPLITCHARDROP=TRUE | FALSE
• CURVELABELSPLITCHAR="*"
Note: The horizontal line to the left of the label represents the maximum end of the
line for reference.
548 Chapter 6 • Plot Statements

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the line label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the line label block.
AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown in
the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which CURVELABELPOSITION=MAX.
Note: The horizontal line to the left of each label represents the maximum end of the
line for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.
LINEPARM Statement 549

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the line.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the line label.

EXTEND=TRUE | FALSE
specifies whether the line is to be drawn to the area bounded by the axes.

Default FALSE

Note If this option is not specified, then there can be a small gap between the
line and the axis. The gap is controlled by the axis offset. If the offset is set
to 0, then there is no gap.

See “boolean ” on page 1339 for other Boolean values that you can use.

GROUP=column | discrete-attr-var | expression

creates a separate parameterized line plot for each unique group value of the
specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of line color and line pattern. Line colors vary
according to the ContrastColor attribute of the GraphData1–
GraphDataN and GraphMissing style elements. The line patterns vary
according to the LineStyle attribute of the GraphData1–GraphDataN
and GraphMissing style elements.

Restriction When this option is used, the X, Y, and SLOPE arguments must
specify numeric columns.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line colors and
line patterns

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Tip The LINEATTRS= option can be used to override the representations

that are used to identify the groups. For example,
LINEATTRS=(PATTERN=SOLID) can be used to assign the same
pattern to all of the lines, letting the line color distinguish group
values. Likewise, LINEATTRS=(COLOR=BLACK) can be used to
550 Chapter 6 • Plot Statements

assign the same color to all of the lines, letting the line pattern
distinguish group values.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

LINEPARM Statement 551

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the line.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor, LineStyle, and LineThickness

attributes of the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Restriction Another plot that establishes a data range for the designed axis must be
included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Restriction Another plot that establishes a data range for the designed axis must be
included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.
552 Chapter 6 • Plot Statements

Details
The LINEPARM statement creates a straight line. You can generate a single line by
specifying a constant for each required argument. You can generate multiple lines by
specifying a numeric column for any or all required arguments. If any of the X= or Y=
columns contains a missing value, then no line is drawn. To request a vertical line,
specify SLOPE=. (specify a missing value as a constant or column value).
A LINEPARM statement can be used in any layout except GRIDDED or OVERLAY3D
layouts. The parent layout must include another plot statement that is derived from data
values that establish a data range for the axes. For example, it can be used with a scatter
plot or a histogram.
To draw vertical or horizontal reference lines, consider using the simpler
REFERENCELINE statement.

Example: LINEPARM Statement

The following graph was generated by the “Example Program” on page 552:

Example Program
The LINEPARM statement draws a line based on a point and the slope of the line that
passes through that point. You can use this statement to create a reference line with any
slope or, in this example, to draw a fit from a linear regression. Many SAS/STAT
procedures create output data sets containing a Y-intercept and slope and coefficient for
the linear regression equation.
proc template;
define statgraph lineparm;
begingraph;
LOESSPLOT Statement 553

entrytitle "Robust Fit of Height and Weight by Sex ";

layout overlay / xaxisopts=(offsetmax=0.35);
scatterplot x=height y=weight / group=sex
markercharacter=eval(substr(sex,1,1))
markercharacterattrs=(size=5pt) datatransparency=0.7;
lineparm x=0 y=intercept slope=slope /
name="Line" group=sex clip=true
curvelabel=eval("Weight = "||put(slope,5.3)||
" * Height + "||put(intercept,6.1))
curvelabellocation=inside
curvelabelattrs=(size=8pt);
discretelegend "Line";
endlayout;
endgraph;
end;
run;

proc sort data=sashelp.heart(keep=height weight sex)

out=heart;
by sex;
run;

ods exclude all;

proc robustreg data=heart method=m
outest=stats(rename=(height=slope));
by sex;
model weight=height;
run;

data all;
merge heart stats(keep=intercept slope sex);
run;

ods select all;

proc sgrender data=all template=lineparm;
run;

LOESSPLOT Statement
Creates a fitted loess curve computed from input data.
Restriction: The LOESSPLOT statement supports only models of one independent and one
dependent variable.
Note: If the input data contains a large number of observations, then it might take several
minutes to generate the plot.
Tips: By default, the LOESSPLOT statement will process up to 5000 observations. If the
input data exceeds 5000 observations, then the plot is not generated. In that case,
you can use the ODS GRAPHICS statement LOESSMAXOBS= option to extend the
limit. See “Details” on page 564.
Starting with the third maintenance release of SAS 9.4, subpixel rendering is enabled
by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page
554 Chapter 6 • Plot Statements

33.For information about the ODS GRAPHICS statement SUBPIXEL= option, see
“ODS GRAPHICS Statement” in SAS ODS Graphics: Procedures Guide.

Syntax
LOESSPLOT X=numeric-column | expression
Y=numeric-column | expression </<regression-options> <option(s)>>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the loess curve.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the loess curve.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string"
specifies a label for the loess curve.
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the curve labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the curve label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the curve label relative to the curve line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
LOESSPLOT Statement 555

specifies whether the split characters are included in the curve label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the curve label block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of curves from just the observations that correspond to
each unique group value of the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the column for the X values.
Y=numeric-column | expression
specifies the column for the Y values.

Optional Arguments
CURVELABEL="string"
specifies a label for the loess curve.

Default No curve label is displayed

Interaction This option is not valid when the GROUP= option is specified.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the curve labels.

Default The GraphValueText style element.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

If the GROUP= option is specified, then this option is ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the curve label relative to the plot area.
556 Chapter 6 • Plot Statements

INSIDE
locates the labels inside the plot area
OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the LOESSPLOT is used in multicell

layouts such as LATTICE, DATAPANEL, or DATALATTICE, where
axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the curve
labels appear.

See “Location and Position of Curve Labels” on page 185

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the curve label relative to the curve line. This option is used
in conjunction with the CURVELABELLOCATION= option to determine where the
curve label appears.
AUTO
automatically positions the curve label near the curve boundary along unused
axes whenever possible (typically Y2 and X2) in order to avoid collision with
tick values.

Restriction This option is used only when

CURVELABELPOSITION=OUTSIDE.

MAX
forces the curve label to appear near maximum curve values (typically, upper
right).
MIN
forces the curve label to appear near minimum curve values (typically, lower
left).
START
forces the curve label to appear near the beginning of the curve.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

END
forces the curve label to appear near the end of the curve.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.
LOESSPLOT Statement 557

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

The AUTO setting is ignored if CURVELABELLOCATION=INSIDE

is specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Note When you specify TICKVALUELIST=, VIEWMAX=, or

VIEWMIN= in an axis statement, the data points that are used to
determine the position of the curve label might fall outside of the
graph area. In that case, the curve label might not be displayed or
might be positioned incorrectly.

See “Location and Position of Curve Labels” on page 185

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters. When a
curve label is split, the label is split on each occurrence of the specified split
characters.

Default FALSE. The curve label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
curve label. In that case, all of the specified split characters together are treated as a
single split character.
When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the curve
label is split unconditionally at each occurrence of any of the specified split
characters. If the curve label does not contain any of the specified characters, then
the label is not split.
"character-list"
one or more characters with no delimiter between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the following
option:
558 Chapter 6 • Plot Statements

curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the curve label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the curve label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.
TRUE
drops the split characters from the curve label text.
FALSE
includes the split characters in the curve label text. When
CURVELABELSPLIT=TRUE and
CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split character, up
to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a curve label with the following
specifications:
• CURVELABELPOSITION=MAX
• CURVELABEL="Product*Group*A"
• CURVELABELSPLIT=TRUE
• CURVELABELSPLITCHARDROP=TRUE | FALSE
• CURVELABELSPLITCHAR="*"
Note: The horizontal line to the left of the label represents the maximum end of the
curve for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the curve label.
LOESSPLOT Statement 559

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the curve label block.
AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown in
the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which CURVELABELPOSITION=MAX.
Note: The horizontal line to the left of each label represents the maximum end of the
curve for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the loess curve.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

GROUP=column | discrete-attr-var | expression

creates a distinct set of curves from just the observations that correspond to each
unique group value of the specified column.
560 Chapter 6 • Plot Statements

discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color and line pattern. Line colors vary
according to the ContrastColor attribute of the GraphData1–
GraphDataN and GraphMissing style elements. Line patterns vary
according to the LineStyle attribute of the GraphData1–GraphDataN
style elements.

Restriction The input data must be sorted by the GROUP= column.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line colors and
line patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Tip The LINEATTRS= option can be used to override the representations

that are used to identify the groups. For example,
LINEATTRS=(PATTERN=SOLID) can be used to assign the same
pattern to all of the loess curves, letting the line color distinguish
group values. Likewise, LINEATTRS=(COLOR=BLACK) can be
used to assign the same color to all of the curves, letting the line
pattern distinguish group values.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of the
GraphData1–GraphDataN style elements.
LOESSPLOT Statement 561

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the loess curve.

Default The GraphFit style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.
562 Chapter 6 • Plot Statements

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles are
automatically included in the data tip information.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Curve")

Default The column label or column name of the column assigned to the role.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles are
automatically included in the data tip information.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.
LOESSPLOT Statement 563

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Loess Regression Options

ALPHA=positive-number
specifies the confidence level to compute.

Default 0.05

Range 0 < number < 1

Tip ALPHA=0.05 represents a 95% confidence level.

CLM="name"
produces confidence limits for a mean predicted value for each observation. The
confidence level is set by the ALPHA= option.

Interaction "name" is a unique name within the template that is case sensitive and
cannot contain spaces. It must be assigned in order for the confidence
limits to be computed. To display confidence limits, you must use this
name as the required argument of a MODELBAND statement.

DEGREE=1 | 2
specifies the degree of the local polynomials to use for each local regression. The
valid values are 1 for local linear fitting or 2 for local quadratic fitting.

Default 1

INTERPOLATION=LINEAR | CUBIC
specifies the degree of the interpolating polynomials used for blending local
polynomial fits at the kd tree vertices.

CUBIC cubic polynomials

LINEAR linear polynomials

Default LINEAR

MAXPOINTS=positive-integer
specifies the maximum number of predicted points generated for the loess curve as
well as confidence limits.

Default 201
564 Chapter 6 • Plot Statements

SMOOTH=AUTO | positive-number
specifies a regression parameter value.

Default AUTO

REWEIGHT=NONE | positive-integer
specifies the number of iterative re-weighting steps to be done. Such iterations are
appropriate when there are outliers in the data or when the error distribution is a
symmetric long-tailed distribution.

Default NONE

WEIGHT=numeric-column
specifies a column in the input data set that contains values to be used as a priori
weights for a loess fit. The values of the weight column must be nonnegative. If an
observation’s weight is zero, negative, or missing, then the observation is deleted
from the analysis.

Details
The LOESSPLOT statement only supports statistical models of one independent and one
dependent variable. For more information about the fitting methodology, see the LOESS
procedure in the SAS/STAT user’s guide.
In addition to the loess curve, the LOESSPLOT statement can compute confidence
levels for the fitted line. To display the confidence levels:
1. Use the CLM= option to declare a name for the confidence level of the mean .
2. Use a MODELBAND statement to refer this name. This statement draws a
confidence band from this information. See “MODELBAND Statement” on page
565 for information about how to control the appearance of the confidence band.

By default, the LOESSPLOT statement will process up to 5000 observations. If the input
data contains more than 5000 observations, then the plot is not drawn and the following
note is written to the SAS log:

NOTE: The number of observations of the LOESS plot (nnnn) exceeds the
limit of 5000. Specify the LOESSMAXOBS option of the ODS GRAPHICS statement to
override the limit.

In that case, you can use the following statement to extend the limit:
ods graphics / loessmaxobs=nnnn

where nnnn is the new limit.

Note: When the input data contains a large number of observations, it might take several
minutes to generate the plot.
For more information about the LOESSMAXOBS= option, see “ODS GRAPHICS
Statement” in SAS ODS Graphics: Procedures Guide.

Example: LOESSPLOT Statement

The following graph was generated by the “Example Program” on page 565:
MODELBAND Statement 565

Example Program
proc template;
define statgraph loessplot;
begingraph;
entrytitle "Loess Fit Plot";
layout overlay;
scatterplot x=weight y=mpg_highway /
datatransparency=0.7;
loessplot x=weight y=mpg_highway / name="fitline"
alpha=0.05 legendlabel="Loess Fit";
discretelegend "fitline";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=loessplot;

run;

MODELBAND Statement
Creates a band showing confidence limits for an associated smoother plot.
Requirement: A MODELBAND statement must be associated with a smoother statement
(LOESSPLOT, REGRESSIONPLOT, or PBSPLINEPLOT) that specifies a fitted
model and a type of confidence level to compute.
Interaction: Starting with the second maintenance release of SAS 9.4, a confidence band that
depicts confidence limits for individual predicted values (CLI) for a weighted spline
plot or regression plot is displayed as a high-low chart instead of a band.
566 Chapter 6 • Plot Statements

Syntax
MODELBAND "confidence-name" </option(s)>;

Summary of Optional Arguments

Appearance options
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.
DATATRANSPARENCY=number
specifies the degree of the transparency of the band fill and the band outline.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display an outlined area, a filled area, or an outlined and
filled modelband area.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled modelband area.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the modelband outlines.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the upper and lower band labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the band labels relative to the plot area.
CURVELABELLOWER="string"
specifies a label for the lower band limit.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the band label relative to the band line.
CURVELABELUPPER="string"
specifies a label for the upper band limit.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.
MODELBAND Statement 567

Required Argument
"confidence-name"
specifies the case-sensitive name assigned to a confidence option in a smoother plot
statement.

Requirement confidence-name must have been assigned to the CLM= or CLI=

option on a smoother plot statement such as LOESSPLOT,
REGRESSIONPLOT, or PBSPLINEPLOT.

Optional Arguments
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
AUTO
specifies that anti-aliasing is controlled by the ANTIALIAS= option in the ODS
GRAPHICS statement.
OFF
specifies that anti-aliasing is always disabled for this plot.

Default AUTO

Interaction This option overrides the ANTIALIAS= option in the ODS

GRAPHICS statement.

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the upper and lower band labels.

Default The GraphValueText style element.

Interactions For this option to take effect, the CURVELABELLOWER= or

CURVELABELUPPER= option must also be specified.

If the smoother statement’s GROUP= option is specified, then this

option is ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOWER="string"
specifies a label for the lower band limit.

Default No curve label is displayed for the lower band

Interaction If the smoother statement’s GROUP= option is specified, then this

option is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELUPPER="string"
specifies a label for the upper band limit.
568 Chapter 6 • Plot Statements

Default No curve label is displayed for the upper band

Interaction If the smoother statement’s GROUP= option is specified, then this

option is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the band labels relative to the plot area.
INSIDE
locates the labels inside the plot area
OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the MODELBAND is used in multi-

cell layouts such as LATTICE, DATAPANEL, or DATALATTICE,
where axes might be external to the grid.

Interactions For this option to take effect, the CURVELABELLOWER= or

CURVELABELUPPER= option must also be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the band
labels appear.

See “Location and Position of Curve Labels” on page 185

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the band label relative to the band line.
AUTO
automatically positions the band labels near the band boundary along unused
axes whenever possible (typically Y2 and X2).

Restriction This option is used only when

CURVELABELPOSITION=OUTSIDE.

MAX
forces the band label to appear near maximum band values (typically, upper
right)
MIN
forces the band label to appear near minimum band values (typically, lower left)
START
forces the band label to appear near the beginning of the band.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

END
forces the band label to appear near the end of the band.
MODELBAND Statement 569

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restrictions The AUTO setting is ignored if CURVELABELLOCATION=INSIDE

is specified.

The START and END settings are ignored if

CURVELABELLOCATION=OUTSIDE is specified.

Interactions For this option to take effect, the CURVELABELLOWER= or

CURVELABELUPPER= option must also be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the band
label appears.

Note When you specify TICKVALUELIST=, VIEWMAX=, or

VIEWMIN= in an axis statement, the data points that are used to
determine the position of the band label might fall outside of the graph
area. In that case, the band label might not be displayed or might be
positioned incorrectly.

See “Location and Position of Curve Labels” on page 185

DATATRANSPARENCY=number
specifies the degree of the transparency of the band fill and the band outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the curve label.

Tip The FILLATTRS= option can be used to set transparency for just the filled
band area. You can combine this option with FILLATTRS= to set one
transparency for the band outline but a different transparency for the band
fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display an outlined area, a filled area, or an outlined and filled
modelband area.
STANDARD
displays a filled band with no outlined
ALL
displays an outlined, filled band
570 Chapter 6 • Plot Statements

(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays an outlined band

FILL displays a filled band

Default The GraphBand:DisplayOpts style reference.

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the band.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled modelband area.

Defaults For non-grouped data, the GraphConfidence:Color style reference.

For grouped data, the GraphData1:Color–GraphDataN:Color style

references.

Interaction For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

Tip The DATATRANSPARENCY= option sets the transparency for band

outline and the band fill. You can combine this option with
DATATRANSPARENCY= to set one transparency for the band outline
but a different transparency for the band fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction The smoother statement’s GROUP= option overrides this option.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the modelband outlines.
MODELBAND Statement 571

Defaults For non-grouped data, the GraphConfidence style element.

For grouped data, the GraphData1: ContrastColor–

GraphDataN:ContrastColor style references.

Interaction If DISPLAY=(FILL), then this option has no effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs. The role-names X,
LIMITLOWER, LIMITUPPER, GROUP, and INDEX are available to indicate
which data tip values to format.

Example TIPFORMAT=(LIMITUPPER=5.3 LIMITLOWER=5.3)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs. The role-names X,
LIMITLOWER, LIMITUPPER, GROUP, and INDEX are available to indicate
which data tip values to label.

Example TIP=(X)
TIPLABEL=(X="Type")

Default The column label or column name of the column assigned to the role.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Requirement The setting for this option should be the same as for the smoother
statement referenced by the confidence-name .

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.
572 Chapter 6 • Plot Statements

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Requirement The setting for this option should be the same as for the smoother
statement referenced by the confidence-name .

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Example: MODELBAND Statement

The following graph was generated by the “Example Program” on page 572:

Example Program
proc template;
define statgraph modelband;
begingraph;
entrytitle "Spline Fit with Confidence Bands";
layout overlay;
modelband "cliband" / name="confband1" display=all
legendlabel="90% CLI" fillattrs=GraphConfidence;
modelband "clmband" / name="confband2" display=all
legendlabel="90% CLM" fillattrs=GraphConfidence2;
scatterplot x=weight y=mpg_highway /
MOSAICPLOTPARM Statement 573

datatransparency=0.7;
pbsplineplot x=weight y=mpg_highway / name="fitline"
clm="clmband" cli="cliband"
alpha=0.1 legendlabel="Spline Fit";
discretelegend "fitline" "confband1" "confband2";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=modelband;

run;

MOSAICPLOTPARM Statement
Creates a mosaic plot from pre-summarized categorical data.
Restriction: You can use the MOSAICPLOTPARM statement in GRIDDED, LATTICE, and
REGION layouts only.
Note: On z/OS hosts, the mosaic plot categories might appear in an order other than data
order.

Syntax
MOSAICPLOTPARM CATEGORY=(column-list)
COUNT=non-negative-numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
COLORGROUP=column | discrete-attr-var
specifies the category column to use for discrete fill colors for the tiles.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies a numeric columnto use to map tile fill colors to a continuous
gradient.
DATATRANSPARENCY=number
specifies the degree of the transparency of the tile fill, outlines, and the values
that are located inside the tiles if those values are displayed.
DISPLAY=STANDARD | ALL | (display-options)
specifies which graphical features to display in the plot.
FILLATTRS=style-element | (fill-options)
specifies the appearance of the tile fill areas.
GUTTER=dimension | (dimension-list)
specifies the gutter (gap) between the splits.
OUTLINEATTRS=style-element | (line-options)
specifies the appearance of the tile outlines.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.
574 Chapter 6 • Plot Statements

SQUARED=TRUE | FALSE
specifies that a square aspect ratio be used for the plot area.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information that is displayed when the cursor is positioned over
a tile.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
INSIDEVALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the category values when they are
located inside a tile.
LABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the category labels.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
VALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the category values that are located
outside of the tiles.
VALUELOCATION=AUTO | INSIDE | OUTSIDE
specifies the location of the category column values in a two-way plot.
XVALUEFITPOLICY=ROTATE | NONE
specifies a policy for avoiding collisions along the width of the plot among
category values that are outside of the tiles.
YVALUEFITPOLICY=NONE | ROTATEALWAYS
specifies a policy for avoiding collisions along the height of the plot among
category values that are outside of the tiles.

ODS options
URL=string-column
specifies the URL of an HTML page to display when a tile is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
CATEGORY=(column-list)
specifies a list of columns of category (classification) values.

Restriction No more than three columns can be specified.

MOSAICPLOTPARM Statement 575

COUNT=non-negative-numeric-column | expression
specifies the column of counts (pre-summarized) for each of the category value
combinations.

Restriction The column values cannot be negative.

Tip You need to provide only the category crossings with nonzero counts.

Optional Arguments
COLORGROUP=column | discrete-attr-var
specifies the category column to use for discrete fill colors for the tiles.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Each unique value of this column is mapped to the COLOR attribute of the
GraphData1–GraphDataN style elements that are in effect. If a discrete attribute map
variable is specified, the color mapping from its associated DISCRETEATTRMAP
statement is used.

Restriction This column or the associated column in the discrete attribute map
variable must be one of the columns in the category column list.

Interactions This option is ignored if the COLORRESPONSE= option is specified.

This option overrides the FILLATTRS= option.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

576 Chapter 6 • Plot Statements

Default The ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to the
color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies a numeric columnto use to map tile fill colors to a continuous gradient.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Interactions When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map.

When only the outlines are displayed, this option overrides suboption
COLOR= in the OUTLINEATTRS= option and varies the outline
color according to the color gradient or the attribute map.

Tip To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

DATATRANSPARENCY=number
specifies the degree of the transparency of the tile fill, outlines, and the values that
are located inside the tiles if those values are displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISPLAY=STANDARD | ALL | (display-options)

specifies which graphical features to display in the plot.
STANDARD
displays tiles with fills, outlines, labels, and values.
ALL
same as STANDARD.
display-options
a space-separated list of one or more of the following options, enclosed in
parentheses:

FILL displays filled tiles.

OUTLINE displays the tile outline.
MOSAICPLOTPARM Statement 577

LABELS displays the category column labels.

TICKS displays the category ticks.
VALUES displays the category values.

Default STANDARD

Interactions If neither FILL nor OUTLINE are present in the display-options list,
then filled and outlined tiles are displayed.

If YVALUELOCATION=INSIDE or if YVALUELOCATION=AUTO
and is effectively set to INSIDE, then the axis ticks are not displayed
even if the display of the ticks is specified for the axis.

FILLATTRS=style-element | (fill-options)
specifies the appearance of the tile fill areas. See “General Syntax for Attribute
Options” on page 1347 for the syntax on using a style element. See “Fill Options”
on page 1348 for the available fill options.

Default The GraphDataDefault style element

Interaction When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is
ignored, and the fill colors vary according to the gradient.

Note When style-element is specified, only the style element’s COLOR

attribute is used.

GUTTER=dimension | (dimension-list)
specifies the gutter (gap) between the splits. The splits occur in the following way:
• When a single dimension is specified, the dimension applies to the gap for the
last split, which has the smallest gap. From the next-to-last last split to the first
split, the gutter is doubled on each split. A single dimension has the effect of
setting a minimum gap for the plot. The following figure shows an example in
which GUTTER=10.

Note: The plot contains equally sized tiles for demonstration purposes.
In this case, the gaps are 30 pixels for the first split, 20 pixels for the second split,
and 10 pixels for the third split (minimum gap).
578 Chapter 6 • Plot Statements

• When a list of dimension values is used, the values apply to each split in the
order in which they are specified. The following figure shows an example in
which GUTTER=(10 20 30).

In this case, the gaps are 10 pixels for the first split, 20 pixels for the second split,
and 30 pixels for the third split.

Default 3px. Dimensions with no units are assumed to be in pixels.

See “dimension” on page 1340

INSIDEVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the category values when they are located
inside a tile. See “General Syntax for Attribute Options” on page 1347for the syntax
on using a style element. See “Text Options” on page 1351 for the available text
options.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all the
font properties such as color, family, size, weight, and style, then the
non-specified properties are derived from the GraphValueText style
element.

Note When style-element is specified, only the element’s COLOR,

FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT
attributes are used.

Tip You can use the VALUEATTRS= option to change the text attributes
for the values that are located outside the plot area.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the category labels. See “General Syntax for
Attribute Options” on page 1347for the syntax on using a style element. See “Text
Options” on page 1351 for the available text options.

Default The GraphLabelText style element.

Interaction If one or more text options are specified and they do not include all the
font properties such as color, family, size, weight, and style, then the
non-specified properties are derived from the GraphLabelText style
element.
MOSAICPLOTPARM Statement 579

Note When style-element is specified, only the style element’s COLOR,

FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT
attributes are used.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the COLORGROUP= option is in effect, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | (line-options)
specifies the appearance of the tile outlines. See “General Syntax for Attribute
Options” on page 1347for the syntax on using a style element. See “Line Options”
on page 1349 for the available line options.

Default The GraphOutlines style element.

Interaction When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the outline colors vary according to the gradient.

Note When style-element is used, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS are used.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either the
ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.
580 Chapter 6 • Plot Statements

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles. The predefined roles are CATEGORY1–
CATEGORYn (in the order in which they are specified in the
CATEGORY=option), COUNT, COLORGROUP, and
COLORRESPONSE.

SQUARED=TRUE | FALSE
specifies that a square aspect ratio be used for the plot area.

Default FALSE

Restriction This option applies to multi-way plots only.

Tip Setting this option to TRUE makes the height of the plot the same as its
width, which can make it easier to compare the proportions.

See “boolean ” on page 1339 for other Boolean values that you can use.

TIP=(role-list) | NONE
specifies the information that is displayed when the cursor is positioned over a tile. If
this option is used, then all of the information that is displayed by default is replaced.
Roles for columns that do not contribute to the bar chart can be specified along with
roles that do.
(role-list)
an ordered, space-separated list of unique MOSAICPLOTPARM roles and user-
defined roles. The MOSAICPLOTPARM roles include COUNT,
COLORGROUP, and COLORRESPONSE. User-defined roles are defined with
the ROLENAME= option.
NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns that are assigned to the category columns and COUNT
roles are automatically included in the data tip information.

Requirement To generate data tips, you must include an ODS GRAPHICS ON

statement with the IMAGEMAP option specified. You must also
write the graphs to the ODS HTML destination.

Interaction The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

Example To display data tips for the columns that are assigned to the roles X
and Y as well as the user-defined role TIP1:
ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
MOSAICPLOTPARM Statement 581

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies the URL of an HTML page to display when a tile is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
tile that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable tiles, you must include an ODS GRAPHICS

ON statement with the IMAGEMAP option specified. You must also
write the graphs to the ODS HTML destination.

Tip The URL value can be blank for some tiles, which means that no
action is taken when those tiles are selected. The URL value can be
the same for different tiles, which means that the same action is taken
when those tiles are selected.

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the category values that are located outside
of the tiles. See “General Syntax for Attribute Options” on page 1347for the syntax
on using a style element. See “Text Options” on page 1351 for the available text
options.
582 Chapter 6 • Plot Statements

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all the
font properties such as color, family, size, weight, and style, then the
non-specified properties are derived from the GraphValueText style
element.

Note When style-element is used, only the style element’s COLOR,

FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT
attributes are used.

Tip You can use the INSIDEVALUEATTRS= option to change text

attributes for values that are located inside the tiles.

VALUELOCATION=AUTO | INSIDE | OUTSIDE

specifies the location of the category column values in a two-way plot.
AUTO
locates the values for the second category in the CATEGORY= list that are
outside of the tiles. If the first category in the CATEGORY= list has any missing
crossings with the second category, or if any of the second category values
collide, then the values are located inside the tiles per the default fit policy.
INSIDE
for each category value, locates the values inside the largest tile.
OUTSIDE
locates the values outside of the tiles, in the plot area.

Default AUTO

Restriction This option applies to two-way plots only.

XVALUEFITPOLICY=ROTATE | NONE
specifies a policy for avoiding collisions along the width of the plot among category
values that are outside of the tiles.
ROTATE
rotates the values if any of the values collide.
NONE
does not attempt to fit values that collide.

Default ROTATE

YVALUEFITPOLICY=NONE | ROTATEALWAYS
specifies a policy for avoiding collisions along the height of the plot among category
values that are outside of the tiles.
NONE
does not attempt to fit values that collide.
ROTATEALWAYS
rotates the values regardless of whether any of the values collide.

Default NONE

Interaction This option is effective only when VALUELOCATION=OUTSIDE.

Example: MOSAICPLOTPARM Statement 583

Details
A mosaic plot displays relative frequencies for categorical variables. Each crossing of
the categorical values is represented by a tile. The area of each tile is proportional to the
frequency of that crossing. The plot is the result of an iterative process. The first iteration
splits the plot area into tiles along the width according to the relative frequency of the
first category column values. Subsequent iterations split the tiles from the previous
iteration in the direction orthogonal to the previous split by using the relative frequencies
of each category column's values. By default, the gap (or gutter) for each split gets
progressively smaller, with a minimum gap of 3 pixels. You can use the GUTTER=
option to specify a different gap for each split.
The following figure provides an example of a three-way mosaic plot, which has three
categories.

Note: The plot contains equally sized tiles for demonstration purposes.
In the example plot, the first split is along the width for CATEGORY 1. The second split
is along the height for CATEGORY 2. Finally, the third split is along the width for
CATEGORY 3. Notice how the gaps between the tiles get progressively smaller from the
first split to the last split.

Example: MOSAICPLOTPARM Statement

The following graph was generated by the “Example Program” on page 584:
584 Chapter 6 • Plot Statements

Example Program
/* Summarize the SASHELP.CARS data for ORIGIN and TYPE */
proc summary data=sashelp.cars nway;
class origin type;
var mpg_highway;
output out=mileage mean=avgMpg N=count / noinherit;
run;

/* Generate the plot */

proc template;
define statgraph mosaicPlotParm;
begingraph;
layout region;
mosaicPlotParm category=(type origin) count=count /
name="mosaic" colorresponse=avgMpg;
continuouslegend "mosaic" / title="Average Miles Per Gallon"
pad=(left=5);
endlayout;
endgraph;
end;
run;

proc sgrender data=mileage template=mosaicPlotParm;

run;

NEEDLEPLOT Statement
Creates a plot of observations as points connected to a baseline by vertical line segments.
NEEDLEPLOT Statement 585

Syntax
NEEDLEPLOT X=column | expression
Y=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing
or bin width.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the needle plot lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the needle lines, markers, and data
labels, if displayed.
DISPLAY=STANDARD | ALL | display-options
specifies whether to display needle lines with or without markers.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line
pattern) to one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the needle lines for the data points.
MARKERATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the data markers.

Axes options
BASELINEINTERCEPT=number | RELATIVE
specifies the Y-intercept for the baseline.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
needle line or marker.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.
586 Chapter 6 • Plot Statements

Label options
DATALABEL=column
specifies labels at the data points.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |
CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the end of the needle lines
and markers, if displayed.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all needle lines and markers from discrete X
values when graphing multiple response variables side by side on a common
axis.
GROUP=column | discrete-attr-var | expression
creates a distinct set of needle lines, markers, and data labels for each unique
group value of the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped needle lines are overlaid or clustered around the
category midpoints on a discrete axis or around the intervals on an interval
axis.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a needle or marker is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies a column or expression for the X values.
NEEDLEPLOT Statement 587

Y=numeric-column | expression
specifies a numeric column or numeric expression for the Y values.

Optional Arguments
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.

Default The GraphAxisLines style element.

Notes The baseline is always drawn by default.

When style-element is specified, only the style element’s COLOR,

LINESTYLE, and LINETHICKNESS attributes are used.

Tip To suppress the baseline, set the line thickness to 0:

baselineattrs=(thickness=0)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

BASELINEINTERCEPT=number | RELATIVE
specifies the Y-intercept for the baseline. The baseline is always displayed in the
chart, whether for a specified value or the default value. When this option is used,
the axis range is adjusted to include the baseline, and the baseline is placed at the
specified value on the response axis.
number
specifies the Y-intercept value to use for the baseline.

Interaction When number is specified, if necessary, the response axis data

range is extended to include the baseline intercept. When a
logarithmic response axis is requested and number is 0 or a negative
value, the response axis reverts to a linear axis. To restore the log
axis in that case, set BASELINEINTERCEPT= to a positive value.

Tips The baseline does not add a tick or a tick value to the axis. To label
the baseline, you can overlay a REFERENCELINE statement with
the same Y value and use its CURVELABEL option.

The appearance of the baseline is controlled by the

BASELINEATTRS= option.

To suppress the baseline, use the BASELINEATTRS= option to set

the line thickness to 0.

RELATIVE
places the baseline at the Y-axis tick mark closest to the minimum of the range
for the needle data points.

Default 0

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or bin
width.
588 Chapter 6 • Plot Statements

Default 0.85

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.

Requirement For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

Interaction When markers are displayed for interval data and

GROUPDISPLAY=CLUSTER and CLUSTERWIDTH= are in
effect, the size of the markers in each cluster might be reduced to no
less than 5 pixels in order to display the cluster within the smallest
effective midpoint space. If you need larger markers in that case, use
the MARKERATTRS= option to specify a larger marker size.

DATALABEL=column
specifies labels at the data points.

Default No data labels are displayed

Note The position of the labels is adjusted to prevent the labels from
overlapping.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interaction For this option to have any effect, the DATALABEL= option must also
be specified.

Note When the DATALABELPOSITION=AUTO option is in effect, in some

cases, the data label font size might be reduced in order to avoid
overlapping labels and markers.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |

CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the end of the needle lines and
markers, if displayed.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters. When set to
TRUE, the data label is split unconditionally at each occurrence of any of the
specified split characters.

Default FALSE. The data labels are not split.

NEEDLEPLOT Statement 589

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
data label. In that case, all of the specified split characters together are treated as a
single split character.
When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is
split unconditionally at each occurrence of any of the specified split characters. If the
data label does not contain any of the specified characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
TRUE
drops the split characters from the data label.
FALSE
includes the split characters in the data label. When DATALABELSPLIT=TRUE
and DATALABELSPLITCHARDROP=FALSE, each split character remains as
590 Chapter 6 • Plot Statements

the last character in the current line. The characters that follow the split character,
up to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a data label with the following
specifications:
• the data label text for this label is Product*Group*A
• DATALABELSPLIT=TRUE
• DATALABELSPLITCHARDROP=TRUE | FALSE
• DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from the

label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains as the
last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.
AUTO
justifies the labels based on the DATALABELPOSITION= option as shown in
the following table.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which DATALABELPOSITION=TOP.
Note: The gray vertical line at the bottom of each label represents the horizontal
center of the text box for reference.
NEEDLEPLOT Statement 591

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of text.

The text box is anchored the same way that the unsplit text is anchored. For example,
if DATALABELPOSITION=TOP, then the bottom center of the text box is
positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the needle plot lines. The following figure shows a
needle plot with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the GraphSkins:DataSkin
style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interaction This option overrides the BEGINGRAPH statement DATASKIN=

option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the needle lines, markers, and data labels,
if displayed.
592 Chapter 6 • Plot Statements

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISCRETEOFFSET=number
specifies an amount to offset all needle lines and markers from discrete X values
when graphing multiple response variables side by side on a common axis.

Default 0 (no offset, all needle lines and markers are centered on the discrete X
values)

Range -0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right. If the parent layout’s axis options
set REVERSE=TRUE, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DISPLAY=STANDARD | ALL | display-options

specifies whether to display needle lines with or without markers.
STANDARD
displays needle lines without markers.
ALL
displays needle lines with markers.
(display-options)
a space-separated list of one or more options enclosed in parentheses. Currently,
only the MARKERS option is supported, which displays needle lines with
markers.

Default STANDARD

Tip Use the MARKERATTRS= and LINEATTRS= options to control the

appearance of the line and markers.

GROUP=column | discrete-attr-var | expression

creates a distinct set of needle lines, markers, and data labels for each unique group
value of the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color, line pattern, and marker symbol. These
vary according to the ContrastColor, LineStyle, and MarkerSymbol
NEEDLEPLOT Statement 593

attributes of the GraphData1–GraphDataN and GraphMissing style

elements.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of marker
symbols, colors, and line patterns.

The marker size is set by the MARKERATTRS= option.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Note The representations that are used to identify the groups can be
overridden. For example, each distinct group value might be
represented by a different line pattern, but the
LINEATTRS=(PATTERN=pattern) option could be used to assign the
same line pattern to all of the plot’s line patterns, letting line color
indicate group values. Likewise, LINEATTRS=(COLOR=color) could
be used to assign the same color to all lines, letting line pattern
indicate group values.

See “DISCRETEATTRVAR Statement” on page 1297

GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped needle lines are overlaid or clustered around the category
midpoints on a discrete axis or around the intervals on an interval axis.
OVERLAY
centers the needle lines for matching category values on the midpoints. The
needle lines in each set of group values are superimposed on each other.
CLUSTER
clusters the needle lines for matching category values around the midpoints. Each
cluster of group values is centered at the midpoint for the category.

Default OVERLAY

Interactions For this option to take effect, the GROUP= option must also be
specified.

For interval data, when markers are displayed and

GROUPDISPLAY=CLUSTER is in effect, the size of the markers in
each cluster might be reduced to no less than 5 pixels in order to
display the cluster within the smallest effective midpoint space. If you
need larger markers in that case, use the MARKERATTRS= option to
specify a larger marker size.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.
594 Chapter 6 • Plot Statements

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
DESCENDING
orders the groups within a category in descending order.

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.

Tips Use the CLUSTERWIDTH= option to control the distance between

the group markers in a cluster.

Use the INDEX= option to alter the default sequence of visual

attributes that is assigned to the groups.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line pattern)
to one of the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.
NEEDLEPLOT Statement 595

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y-column label. If a label is not defined, then the Y-column name
is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the needle lines for the data points.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor, LineStyle and LineThickness

attributes of the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

MARKERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes of

the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions This option’s COLOR= suboption overrides the default behavior for
grouped data. When the COLOR= suboption is specified in that case,
596 Chapter 6 • Plot Statements

all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior for

grouped data. When the SYMBOL= suboption is specified in that
case, all markers have the same symbol, and the symbol color alone
distinguishes the markers.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

This option is ignored if the DISPLAY= option disables the display of

the markers.

If the DATASKIN= option is in effect, then the data skin determines

the marker outlines. Any outline-related settings from the current ODS
style or from the marker attribute options are ignored.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Marker Options” on page 1350 for available marker-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.
NEEDLEPLOT Statement 597

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X. Y, DATALABEL, INDEX, and GROUP.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a needle line
or marker. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the needle plot can
be specified along with roles that do.
(role-list)
an ordered, space-separated list of unique NEEDLEPLOT and user-defined roles.
NEEDLEPLOT roles include X, Y, DATALABEL, and GROUP.
User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns assigned to
the roles X and Y as well as the column Obs, which is not assigned to
any pre-defined NEEDLEPLOT role. The Obs column must first be
assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X, Y, DATALABEL, and GROUP.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.
598 Chapter 6 • Plot Statements

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a needle or marker is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
needle that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable needle lines, you must include an ODS

GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
Example: NEEDLEPLOT Statement 599

specified in the LINEAROPTS= or TIMEOPTS= option for either

axis.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding needle or marker is
selected.

The URL value can be the same for any X and Y pairs. In that case,
the same action is taken when the needle or marker is selected for
those X and Y pairs.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
In the NEEDLEPLOT statement, the X column can specify character or numeric values.
The Y column must specify numeric values. For character columns, the X-axis is always
of TYPE=DISCRETE. For numeric columns, the X-axis is of TYPE=LINEAR by
default.
The Y-axis is of TYPE=LINEAR by default.

Example: NEEDLEPLOT Statement

The following graph was generated by the “Example Program” on page 600:
600 Chapter 6 • Plot Statements

Example Program
proc template;
define statgraph needleplot;
begingraph;
entrytitle "IBM Stock Trend";
layout overlay;
needleplot x=date y=close /
baselineintercept=80 lineattrs=(color=blue);
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.stocks template=needleplot;

where stock="IBM" and date > "31dec1999"d;
run;

PBSPLINEPLOT Statement
Creates a fitted penalized B-spline curve computed from input data.
Restriction: The PBSPLINEPLOT statement supports only models of one independent and one
dependent variable.
Tip: Starting with the third maintenance release of SAS 9.4, subpixel rendering is enabled
by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page
33.For information about the ODS GRAPHICS statement SUBPIXEL= option, see
“ODS GRAPHICS Statement” in SAS ODS Graphics: Procedures Guide.
PBSPLINEPLOT Statement 601

Syntax
PBSPLINEPLOT X=numeric-column | expression
Y=numeric-column | expression </<regression-option(s)> <option(s)>>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the curve and curve label.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line attributes of the spline curve.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string"
specifies a label for the spline curve.
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the spline curve labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the spline curve label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the spline curve label relative to the curve line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the curve label block.
602 Chapter 6 • Plot Statements

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of curves from just the observations that correspond to
each unique group value of the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the column for the X values.
Y=numeric-column | expression
specifies the column for the Y values.

Optional Arguments
CURVELABEL="string"
specifies a label for the spline curve.

Default No curve label is displayed

Interaction This option is not valid when the GROUP= option is specified.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the spline curve labels.

Default The GraphValueText style element.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

If the GROUP= option is specified, then this option is ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the spline curve label relative to the plot area.
INSIDE
locates the labels inside the plot area
PBSPLINEPLOT Statement 603

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the PBSPLINEPLOT is used in

multicell layouts such as LATTICE, DATAPANEL, or
DATALATTICE, where axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the curve
labels appear.

See “Location and Position of Curve Labels” on page 185

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the spline curve label relative to the curve line.
AUTO
automatically positions the spline curve label near the curve boundary along
unused axes whenever possible (typically Y2 and X2) in order to avoid collision
with tick values.

Restriction This option is used only when

CURVELABELPOSITION=OUTSIDE.

MAX
forces the spline curve label to appear near maximum curve values (typically,
upper right)
MIN
forces the spline curve label to appear near minimum curve values (typically,
lower left)
START
forces the spline curve label to appear near the beginning of the curve.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

END
forces the spline curve label to appear near the end of the curve.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

604 Chapter 6 • Plot Statements

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

The AUTO setting is ignored if CURVELABELLOCATION=INSIDE

is specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the spline
curve label appears.

Note When you specify TICKVALUELIST=, VIEWMAX=, or

VIEWMIN= in an axis statement, the data points that are used to
determine the position of the spline curve label might fall outside of
the graph area. In that case, the spline curve label might not be
displayed or might be positioned incorrectly.

See “Location and Position of Curve Labels” on page 185

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters. When a
curve label is split, the label is split on each occurrence of the specified split
characters.

Default FALSE. The curve label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
curve label. In that case, all of the specified split characters together are treated as a
single split character.
When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the curve
label is split unconditionally at each occurrence of any of the specified split
characters. If the curve label does not contain any of the specified characters, then
the label is not split.
"character-list"
one or more characters with no delimiter between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the following
option:
PBSPLINEPLOT Statement 605

curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the curve label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the curve label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.
TRUE
drops the split characters from the curve label text.
FALSE
includes the split characters in the curve label text. When
CURVELABELSPLIT=TRUE and
CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split character, up
to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a curve label with the following
specifications:
• CURVELABELPOSITION=MAX
• CURVELABEL="Product*Group*A"
• CURVELABELSPLIT=TRUE
• CURVELABELSPLITCHARDROP=TRUE | FALSE
• CURVELABELSPLITCHAR="*"
Note: The horizontal line to the left of the label represents the maximum end of the
curve for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the curve label.
606 Chapter 6 • Plot Statements

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the curve label block.
AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown in
the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which CURVELABELPOSITION=MAX.
Note: The horizontal line to the left of each label represents the maximum end of the
curve for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the curve and curve label.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

GROUP=column | discrete-attr-var | expression

creates a distinct set of curves from just the observations that correspond to each
unique group value of the specified column.
PBSPLINEPLOT Statement 607

discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color and line pattern. Line colors vary
according to the ContrastColor attribute of the GraphData1–
GraphDataN and GraphMissing style elements. Line patterns vary
according to the LineStyle attribute of the GraphData1–GraphDataN
and GraphMissing style elements.

Restriction The input data must be sorted by the GROUP= column.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line colors and
line patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Tip The LINEATTRS= option can be used to override the representations

that are used to identify the groups. For example,
LINEATTRS=(PATTERN=SOLID) can be used to assign the same
pattern to all of the loess curves, letting the line color distinguish
group values. Likewise, LINEATTRS=(COLOR=BLACK) can be
used to assign the same color to all of the curves, letting the line
pattern distinguish group values.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of the
GraphData1–GraphDataN style elements.
608 Chapter 6 • Plot Statements

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes of the spline curve.

Default The GraphFit style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.
PBSPLINEPLOT Statement 609

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles are
automatically included in the data tip information.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Curve")

Default The column label or column name of the column assigned to the role.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles are
automatically included in the data tip information.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.
610 Chapter 6 • Plot Statements

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

PBSPLINE Regression Options

ALPHA=positive-number
specifies the confidence level to compute.

Default 0.05

Range 0 < positive-number < 1

Tip ALPHA=0.05 represents a 95% confidence level.

CLI="name"
produces confidence limits for individual predicted values for each observation.

Interaction name is a unique name within the template that is case sensitive and
cannot contain spaces. It must be assigned in order for the confidence
limits to be computed. To display confidence limits, you must use this
name as the required argument of a MODELBAND statement. See the
example in the section “Example: PBSPLINEPLOT Statement” on
page 612 .

CLM="name"
produces confidence limits for a mean predicted value for each observation.

Interaction name is a unique name within the template that is case sensitive and
cannot contain spaces. It must be assigned in order for the confidence
limits to be computed. To display confidence limits, you must use this
name as the required argument of a MODELBAND statement. See the
example in the section “Example: PBSPLINEPLOT Statement” on
page 612 .

DEGREE=non-negative-integer
specifies the degree of B-spline.

Default 3

Ranges 0–174 in the first maintenance release of SAS 9.4 and earlier releases.

0–10 starting with the second maintenance release of SAS 9.4.

PBSPLINEPLOT Statement 611

Restriction Starting with the second maintenance release of SAS 9.4, DEGREE=
and NKNOTS= cannot be set to 0 simultaneously. When both are set to
0, an error results.

FREQ=numeric-column
specifies a column in the input data set that represents the frequency of occurrence of
the current observation, essentially treating the data set as if each observation
appeared n times, where n is the value of the FREQ column for the observation.
Noninteger values of the FREQ column are truncated to the largest integer less than
the FREQ value. The observation is used in the analysis only if the value of the
FREQ column is greater than or equal to 1.
MAXPOINTS=positive-integer
specifies the maximum number of predicted points generated for the spline curve as
well as any confidence limits.

Default 201

NKNOTS=non-negative-integer
specifies the number of evenly spaced internal knots. By default, a large number of
knots (100) is specified, which allows for an extreme lack of smoothness in the
results. However, the final function is typically much smoother due to the penalty.
When SMOOTH=0 is specified, you should typically ask for many fewer knots than
the default, since there is no penalty for lack of smoothness. For example, ten or
fewer knots is usually enough to follow the functional form found in most data.

Default 100

Restriction Starting with the second maintenance release of SAS 9.4, MKNOTS=
and DEGREE= cannot be set to 0 simultaneously. When both are set to
0, an error results.

See example “Penalized B-Splines” in the TRANSREG procedure

description in SAS/STAT User's Guide.

SMOOTH=AUTO | non-negative-number
specifies a regression parameter value.

Default AUTO

Note With SMOOTH=AUTO, a regression parameter that minimizes a lack-of-

smoothness penalty is automatically selected.

Tip You can specify SMOOTH=0 to get an ordinary B-spline fit.

WEIGHT=numeric-column
specifies a column in the input data set that contains values to be used as a priori
weights for a penalized B-spline fit. If an observation’s weight is zero, negative, or
missing, then the observation is deleted from the analysis.

Interaction Starting with the second maintenance release of SAS 9.4, when the
CLI= option is used with this option, the confidence band for individual
predicted values is displayed as a high-low chart instead of a band.
612 Chapter 6 • Plot Statements

Details
The PBSPLINEPLOT statement only supports models of one independent and one
dependent variable. For more information about the fitting methodology, see the
TRANSREG procedure in the SAS/STAT user’s guide.
In addition to the penalized B-spline, the PBSPLINEPLOT statement can compute
confidence levels for the fitted line. To display the confidence levels,
1. use the CLI= or CLM= option to declare a name for the confidence level
2. use a MODELBAND statement to refer to this name. This statement draws a
confidence band from this information. See “MODELBAND Statement” on page
565 for information about how to control the appearance of the confidence band.

Example: PBSPLINEPLOT Statement

The following graph was generated by the “Example Program” on page 612:

Example Program
proc template;
define statgraph pbsplineplot;
begingraph;
entrytitle "Spline Fit";
layout overlay;
scatterplot x=weight y=mpg_highway /
datatransparency=0.7;
pbsplineplot x=weight y=mpg_highway / name="fitline"
alpha=0.05 legendlabel="Spline Fit";
PIECHART Statement 613

discretelegend "fitline";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=pbsplineplot;

run;

PIECHART Statement
Creates a pie chart that is computed from input data.
Requirement: The PIECHART statement must be placed in a LAYOUT REGION, LAYOUT
GRIDDED, or LAYOUT LATTICE block. It cannot be placed in an overlay-type layout
such as LAYOUT OVERLAY or LAYOUT OVERLAYEQUATED because a pie chart
does not have axes.
Note: The PIECHART statement does not honor the ODS GRAPHICS options
DISCRETEMAX=, GROUPMAX=, and LABELMAX=.

Syntax
PIECHART CATEGORY=column | discrete-attr-var | expression </option(s)>;
PIECHART CATEGORY=column | discrete-attr-var | expression
RESPONSE=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
CATEGORYDIRECTION=COUNTERCLOCKWISE | CLOCKWISE
specifies whether to display the pie slices in counterclockwise or clockwise
sequence.
CENTERFIRSTSLICE=TRUE | FALSE
specifies whether the first pie slice is centered on the starting angle or starts
on the starting angle.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled pie slices.
DATATRANSPARENCY=number
specifies the degree of the transparency of all pie slices, outlines, and text.
DISPLAY=STANDARD | (display-options)
specifies whether to display outlined pie slices, filled pie slices, or outlined
and filled pie slices.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled pie slices.
OTHERSLICEOPTS=(other-slice-options)
specifies the properties of the Other slice.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the properties of the pie and slice outlines.
START=degrees
614 Chapter 6 • Plot Statements

specifies which degree between 0 and 360 serves as the starting position for
the first pie slice.

Data tip options

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a pie
slice.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Grouping options
GROUP=column | expression
creates a separate concentric annulus (or stacked cylinders) for each unique
group value of the specified column.
GROUPGAP=dimension
specifies a dimension for the optional gap that can be displayed between each
annulus of a grouped pie.
GROUPLABELOPTS=(grouplabel-options)
specifies text attributes, location, and other options for displaying group
labels.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the Group column are included in the pie.
OTHERSLICE=TRUE | FALSE
specifies whether to consolidate smaller pie slices into a single slice that
represents “other” values that are in the data, or whether to display those
smaller slices as separate pie slices.

Label options
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the slice labels.
DATALABELCONTENT=ALL | STANDARD | NONE | (content-options)
specifies the information to display in the slice labels.
DATALABELLOCATION=AUTO | INSIDE | OUTSIDE | CALLOUT
specifies whether to display the slice labels within the pie slices or outside of
the pie circumference.
LABELFITPOLICY=NONE | DROP
specifies the label fitting policy to be used if a particular label does not fit
within the pie slice.

ODS options
URL=string-column
specifies an HTML page to display when a pie slice is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
STAT=FREQ | PCT | SUM | MEAN
PIECHART Statement 615

specifies the statistic to be computed.

Required Argument
CATEGORY=column | discrete-attr-var | expression
specifies the column for the category values. Duplicated values of CATEGORY are
summarized into a unique value. All values are treated as discrete.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

See “DISCRETEATTRVAR Statement” on page 1297

Optional Arguments
CATEGORYDIRECTION=COUNTERCLOCKWISE | CLOCKWISE
specifies whether to display the pie slices in counterclockwise or clockwise
sequence.

Default COUNTERCLOCKWISE

Tip The START= option controls the starting angle for the first pie slice.

CENTERFIRSTSLICE=TRUE | FALSE
specifies whether the first pie slice is centered on the starting angle or starts on the
starting angle.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
The following figure shows the effect of this option on a pie chart in which Asia is
the first category slice, the starting angle is 0 degrees, and the category direction is
counterclockwise.

Default FALSE

Tips Use the START= option to change the starting angle.

616 Chapter 6 • Plot Statements

Use the CATEGORYDIRECTION= option to change the category

direction.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the slice labels.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

DATALABELCONTENT=ALL | STANDARD | NONE | (content-options)

specifies the information to display in the slice labels.
ALL
displays all available information
STANDARD
equivalent to specifying the two content-options CATEGORY and RESPONSE
NONE
does not display slice labels
(content-options)
a space-separated list of one or more of the following options enclosed in
parentheses:
CATEGORY
displays the CATEGORY value
PERCENT
displays the following based on the setting for the STAT= option:
• when STAT=FREQ or STAT=PCT, the PERCENT value
• when STAT=MEAN or STAT=SUM, nothing
RESPONSE
displays the statistic that is requested in the STAT= option.

Defaults When STAT=PCT, the default is (CATEGORY PERCENT).

Otherwise, the default is STANDARD.

Note The position of the labels is adjusted to prevent the labels from
overlapping.

DATALABELLOCATION=AUTO | INSIDE | OUTSIDE | CALLOUT

specifies whether to display the slice labels within the pie slices or outside of the pie
circumference.
PIECHART Statement 617

AUTO
automatically selects either INSIDE, OUTSIDE, or CALLOUT to optimize the
label position
INSIDE
locates the slice labels inside the pie slices.
Note: If a particular label does not fit within the pie slice, then the fit policy
takes effect (set by the LABELFITPOLICY= option).
OUTSIDE
locates the slice labels outside of the pie circumference.
CALLOUT
locates the slice labels outside of the pie circumference and draws a line from the
label to its slice.

Default AUTO

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled pie slices. The following figure shows
pies with each of the skins applied.
618 Chapter 6 • Plot Statements

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, DISPLAY= FILL must be in effect.
Otherwise, this option is ignored.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The appearance of the data skin is based on the FILLATTRS= color.

When a data skin is applied, all slice outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of all pie slices, outlines, and text.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the pie
slices. The OTHERSLICEOPTS= option can be used to specify
transparency for the “other” slice. You can combine this option with
FILLATTRS= and with OTHERSLICEOPTS= to set one transparency for
the outlines and text but a different transparency for the pie slices.
Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
PIECHART Statement 619

DISPLAY=STANDARD | (display-options)
specifies whether to display outlined pie slices, filled pie slices, or outlined and filled
pie slices.
STANDARD
displays outlined, filled pie slices
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:
OUTLINE
displays outlined pie slices. The default outline properties are set by the
GraphOutline style element.
FILL
displays filled pie slices with each slice a different color. The default colors
are set by the Color attribute of the GraphData1–GraphDataN style elements.
The fill color of the “other” slice (if shown) is from the color attribute of the
GraphOther style element. If FILL is not specified, then an opaque pie is
drawn using the background color of the containing layout.

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the pie slices.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled pie slices. Prior to the third maintenance release
of SAS 9.4, this option specifies the color and transparency of all of the pie slices,
excluding the Other slice. Starting with the third maintenance release of SAS 9.4,
this option specifies the transparency for the Other slice as well.

Default The GraphDataDefault:Color style reference.

Interaction For this option to have any effect, the fill must be enabled by the ODS
style or the DISPLAY= option.

Tips The FILLATTRS= suboption of the OTHERSLICEOPTS= option

specifies the color and transparency of the Other slice.

The DATATRANSPARENCY= option sets the transparency for all pie

slices, outlines, and text. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines and
text but a different transparency for the pie slices. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element value.

“Fill Options” on page 1348 for available fill-options values.

GROUP=column | expression
creates a separate concentric annulus (or stacked cylinders) for each unique group
value of the specified column. The grouped rings are displayed in data order.
620 Chapter 6 • Plot Statements

Interactions When this option is used, the unique column values are found and then
the slice colors are taken from the GraphData1–GraphDataN style
elements.

Missing values in the data can affect the group order. You can use the
INCLUDEMISSINGGROUP= option to manage missing group
values. In addition, you can use
INCLUDEMISSINGDISCRETE=TRUE in the BEGINGRAPH
statement to create pie slices for missing CATEGORY values.

Tip This option creates only rings of pies. To create a grid of pies, specify
the PIECHART statement within a LAYOUT LATTICE, LAYOUT
DATALATTICE, or LAYOUT DATAPANEL statements.

See FILLATTRS=

GROUPGAP=

GROUPLABELOPTS=

GROUPGAP=dimension
specifies a dimension for the optional gap that can be displayed between each
annulus of a grouped pie.

Default 0

Restriction For this option to take effect, the GROUP= option must also be
specified.

Interaction If the specified dimension is too large for the area that is available to
the pie chart, then the results might be unexpected.

Note The size of the inner pie remains the same regardless of the
GROUPGAP= value.
PIECHART Statement 621

See “dimension” on page 1340

GROUPLABELOPTS=(grouplabel-options)
specifies text attributes, location, and other options for displaying group labels. The
following grouplabel-options are available. One or more options can be specified as
space-separated name = value pairs.
LABEL=AUTO | NONE | "string"
specifies a descriptive label for the Group column

AUTO specifies the column label of the GROUP= column or the column
name of the GROUP= column, if no column label exists.
NONE specifies that no label is displayed
"string" specifies a string to use as the label

Default AUTO

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the group label.

Default The GraphLabelText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LOCATION=RIGHT | LEFT
specifies whether the block of text for group labeling appears to the right or left
of the pie.

Default RIGHT

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the group values.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

Default The group label and values are shown as a block of text to the right or
left of the pie. Slice labels are moved to the inside of the pie slices. A
line is drawn from each group value to its annulus (or cylinder).

Restriction For this option to take effect, the GROUP= option must also be
specified.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the Group column are included in the pie.

Default TRUE
622 Chapter 6 • Plot Statements

Restriction For this option to take effect, the GROUP= option must also be
specified.

See “boolean ” on page 1339 for other Boolean values that you can use.

LABELFITPOLICY=NONE | DROP
specifies the label fitting policy to be used if a particular label does not fit within the
pie slice.
NONE
draws each label regardless of whether it fits within the slice region.
DROP
drops labels that do not fit within the slice region, but draws labels that do fit.

Default NONE

Tip This option determines how labels are managed when

DATALABELLOCATION= INSIDE and a particular label does not fit
within the pie slice.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

OTHERSLICE=TRUE | FALSE
specifies whether to consolidate smaller pie slices into a single slice that represents
“other” values that are in the data, or whether to display those smaller slices as
separate pie slices. If this option is set to FALSE, then all unique category values
appear as slices. If this option is set to TRUE, then some of the smaller slices might
be combined into a single slice, referred to as the Other slice.

Default TRUE

Tip To set the properties of the “other” slice, use the OTHERSLICEOPTS=
option.

See “boolean ” on page 1339 for other Boolean values that you can use.

OTHERSLICEOPTS=(other-slice-options)
specifies the properties of the Other slice. Example:
piechart category=region / name="p"
datalabelcontent=(percent) datalabellocation=inside
otherslice=true
othersliceopts=(type=percent percent=11 label="Other Regions") ;
PIECHART Statement 623

The following other-slice-options values are available. You can specify one or more
options as space-separated name = value pairs.
TYPE=PERCENT | MAXSLICES
specifies which method to use to determine the size of the Other slice.

PERCENT uses the percentage that is set by the PERCENT=

suboption.
MAXSLICES uses the count that is set by the MAXSLICES= suboption.

Default PERCENT

MAXSLICES=positive-integer
specifies the maximum number of category values to represent with pie slices.
Any remaining values are consolidated into the Other slice.

Default 10

Interactions For this option to have any effect, TYPE=MAXSLICES must also
be specified among the suboptions for OTHERSLICEOPTS=.

The slices are counted in the order in which they are displayed.
This order is affected by the CATEGORYDIRECTION= option.

PERCENT=percent-of-total
collects all category values with response values less than or equal to the
specified percent-of-total value into the Other slice.

Default 4. Any original slice that represents 4% or less of the total is put in
the Other category.

Range 0–100

Interaction For this option to have any effect, TYPE=PERCENT must also be
specified among the suboptions for OTHERSLICEOPTS=.
624 Chapter 6 • Plot Statements

LABEL="string"
specifies a label for the Other slice.

Default "OTHER"

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the area fill for the Other slice. This option does not
affect the appearance of the area fill for the remaining slices.

Default The GraphOther style element.

Interactions Suboption TRANSPARENCY= in this option overrides the

DATATRANSPARENCY= option for the other slice only.

Starting with the third maintenance release of SAS 9.4, suboption

TRANSPARENCY= in this option overrides suboption
TRANSPARENCY= in option FILLATTRS= in the PIECHART
statement only for the Other slice.

See “General Syntax for Attribute Options” on page 1347 for the
syntax for using a style-element value.

“Fill Options” on page 1348 for available fill-options values.

Interaction This option is ignored if OTHERSLICE=FALSE.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the properties of the pie and slice outlines.

Default The GraphOutlines style element.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Line Options” on page 1349 for available line-options.

START=degrees
specifies which degree between 0 and 360 serves as the starting position for the first
pie slice. A value of 0 degrees corresponds to the three o'clock position. Degrees can
be either positive or negative. Positive values move the starting position
counterclockwise, and negative values move the starting position clockwise. From
the starting point specified by this option, the slices are drawn in the direction
specified by the CATEGORYDIRECTION= option.

Default 0

Range 0 to 360

STAT=FREQ | PCT | SUM | MEAN

specifies the statistic to be computed. For pie charts with no RESPONSE= column:

FREQ frequency count

PIECHART Statement 625

PCT percentages between 0 and 100

For pie charts with a RESPONSE= column:

SUM
MEAN

Defaults SUM for pie charts that specify the RESPONSE= argument

FREQ for pie charts that do not specify the RESPONSE= argument

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a pie slice. If
this option is used, then it replaces all of the information that is displayed by default.
(role-list)
an ordered, space-separated list of unique PIECHART roles. PIECHART roles
include CATEGORY, RESPONSE, and GROUP. The RESPONSE role
represents the computed statistic for the CATEGORY value, based on the statistic
that is set by the STAT= option.

Example The following example displays data tips for the columns assigned to
the roles CATEGORY and RESPONSE.
TIP=(CATEGORY RESPONSE)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: CATEGORY and RESPONSE .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIP=(RESPONSE)
TIPFORMAT=(RESPONSE=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

626 Chapter 6 • Plot Statements

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example TIP=(RESPONSE)
TIPLABEL=(RESPONSE="Average Sales")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

URL=string-column
specifies an HTML page to display when a pie slice is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
pie slice that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Restriction A generated Other slice does not have a URL. See OTHERSLICE=.

Requirement To generate a plot with selectable pie slices, you must include an
ODS GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option has no effect when TIP=NONE.

Tips The URL value can be blank for some pie slices, meaning that no
action is taken when the corresponding slice is selected.

The URL value can be the same for any CATEGORY and
RESPONSE pairs. In that case, the same action is taken when the pie
slices for those pairs are selected.

Optional Response Argument

RESPONSE=numeric-column | expression
specifies response values that are read from a numeric column or an expression.

Details
The input data for the PIECHART statement is raw, unsummarized input data. The
PIECHART performs discrete binning for the Category column and calculates
appropriate summarization statistics (sum, mean, and so on) based on the setting for the
STAT= option.
The fill color of each pie slice is derived from the Color attribute of the GraphData1–
GraphDataN style elements as described in “Attribute Rotation Patterns” in SAS Graph
Example: PIECHART Statement 627

Template Language: User's Guide. The default order of the pie slices depends on the
data type of the CATEGORY values:
• For numeric data, the slices appear in the ascending order of the unformatted data
values.
• For discrete data, the slices appear in data order.
You can use the START= and CATEGORYDIRECTION= options to control the pie slice
positions and display order.
By default, the pie slices are labeled with the CATEGORY and RESPONSE values,
which are displayed inside the slices. You can use the DATALABELLOCATION= and
DATALABELCONTENT= options to control where the pie slices are labeled and the
label content.
By default, if two or more slices take up less than 4% of a pie, then an “other” slice is
created by consolidating those small slices. To change the default criteria, use the
OTHERSLICE= and OTHERSLICEOPTS= options. The calculated “other” slice is
displayed as the last slice in the pie, and as the last legend entry for the pie. If a category
value is the same as the “other” slice label, then two slices might be displayed with the
same label ("Other" by default) and different fill attributes. In that case, both slices are
represented in the pie legend.
To create a pie slice for missing CATEGORY values, specify
INCLUDEMISSINGDISCRETE=TRUE in the BEGINGRAPH statement. The fill color
of the missing category slice is assigned the fill color from the GraphMissing style
element except when a user-defined format is applied to the category value. In that case,
the missing category slice is assigned the fill color from a GraphData1–GraphdDataN
style element in data order instead.
Note: The PIECHART statement does not honor the MISSING= system option.
Regardless of the MISSING= system option value, unless a user-defined format is
applied to the value, the default missing-numeric-value character (.) is used to depict
missing numeric values.

Example: PIECHART Statement

The following graph was generated by the “Example Program” on page 628 :
628 Chapter 6 • Plot Statements

Example Program
proc template;
define statgraph simplepie;
begingraph;
entrytitle "Car Models by Origin";
layout region;
piechart category=origin / datalabellocation=outside;
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.cars
template=simplepie;
run;

POLYGONPLOT Statement
Draws a polygon from data that is stored in a data set.
Note: This statement is valid in the first maintenance release of SAS 9.4 and later
releases.
Tip: Starting with the third maintenance release of SAS 9.4, you can use subpixel
rendering with this statement. It is enabled by default. To disable subpixel rendering,
specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 33. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.
POLYGONPLOT Statement 629

Syntax
POLYGONPLOT X=column | expression Y=column | expression
ID=column | expression< /options(s)>

Summary of Optional Arguments

Appearance options
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.
BACKLIGHT=number | AUTO
specifies a back-light effect for the polygon label text.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the numeric column or range attribute map variable to use to
determine the polygon colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of filled polygons.
DATATRANSPARENCY=number
specifies the degree of the transparency of the polygon fill, outline, and label,
when these attributes are displayed.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display the polygon outline, fill, or both.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled polygon areas.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the polygon outline.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.
ROTATE=numeric-column | numeric-constant | expression
specifies the angle of rotation for the polygon, measured in degrees.
ROTATELABEL=AUTO | NONE | VERTICAL
specifies the rotation of the polygon label with respect to the rotation of the
polygon.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
630 Chapter 6 • Plot Statements

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
polygon.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
LABEL=column | expression
specifies the label for the polygon.
LABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the polygon label.
LABELLOCATION=INSIDEBBOX | OUTSIDEBBOX | OUTSIDE
specifies the location of the polygon label.
LABELPOSITION=CENTER | XMIN | XMAX | YMIN | YMAX
specifies the position of the polygon label with respect to the label location.
LABELSPLIT=TRUE | FALSE
specifies whether to split the polygon label at the specified split characters.
LABELSPLITCHAR="character-list"
specifies one or more characters on which the polygon label can be split if
needed.
LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the polygon label.
LABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the polygon label
blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies the amount by which to offset all polygon vertices from discrete X
values, from discrete Y values, or from both.
GROUP=column | discrete-attr-var | expression
creates a separate and visually distinctive polygon for each unique grouping
value.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.
XOFFSET=numeric-column | expression
specifies an individual offset from the discrete X value on each vertex of the
polygon.
YOFFSET=numeric-column | expression
specifies an individual offset from the discrete Y value on each vertex of the
polygon.

ODS options
URL=string-column
specifies an HTML page that is displayed when the polygon is selected.

Plot reference options

POLYGONPLOT Statement 631

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies the column for the X values.

Note A missing value in the X column signals the start of the data for a hole in the
polygon. See “Drawing Holes in a Polygon” on page 650.

Y=column | expression
specifies the column for the Y values.

Note A missing value in the Y column signals the start of the data for a hole in the
polygon. See “Drawing Holes in a Polygon” on page 650.

ID=column | expression
specifies the column that contains the ID value that is associated with each polygon.

Restriction Only unformatted values in the ID= column are used.

Requirements All of the observations for a single polygon must have the same ID
value.

When multiple polygons are defined in the same data set, all of the
observations for a given ID must be defined contiguously.
Interspersing the ID observations in the data set can produce
unexpected results.

Note Observations that have a missing value in the ID column are ignored
by the POLYGONPLOT statement.

Optional Arguments
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
AUTO
specifies that anti-aliasing is controlled by the ANTIALIAS= option in the ODS
GRAPHICS statement.
OFF
specifies that anti-aliasing is always disabled for this plot.

Default AUTO

Interaction This option overrides the ANTIALIAS= option in the ODS

GRAPHICS statement.

BACKLIGHT=number | AUTO
specifies a back-light effect for the polygon label text.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
632 Chapter 6 • Plot Statements

The effect is applied only to the polygon label text.

number
specifies the degree of the back-light effect.

Range 0–1, where 0 specifies no effect and 1 specifies maximum effect

AUTO
the system selects an appropriate level for the back-light effect. If the GROUP=
or COLORRESPONSE= option is in effect, BACKLIGHT=0.75. Otherwise,
BACKLIGHT=0.5.
The following figure shows the effect on a polygon label located inside the polygon
bounding box.

The back light is based on text color. For dark colors, a contrasting white back-light
effect is used. For lighter colors, a contrasting black back-light effect is used. The
following figure shows the two back-light types when BACKLIGHT=1.

Default 0 (no back-light effect)

Restriction Vector graphics output cannot be generated when the back-light effect
is applied. If you request vector graphics output and enable the back-
light effect, an image is generated instead.

Interaction The LABEL= option must be specified for this option to have any
effect.

Tip The BACKLIGHT= option is most effective when the text color has a
low level of contrast with the background or when the background is
cluttered.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
POLYGONPLOT Statement 633

ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Defaults For a filled polygon, the ThreeColorRamp style element

For an unfilled polygon, the ThreeColorAltRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to the
color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the numeric column or range attribute map variable to use to determine the
polygon colors.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Requirement The COLORRESPONSE= value should remain constant for the same
ID value. Otherwise, unexpected results might occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the GROUP= option is ignored.

When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map.

When only the outline is displayed, this option overrides suboption

COLOR= in the OUTLINEATTRS= option and varies the outline
color according to the color gradient or the attribute map.

Tip To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.
634 Chapter 6 • Plot Statements

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of filled polygons. The following figure shows a
polygon with each of the skins applied.

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the DISPLAY= option must
include FILL.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The appearance of the data skin is based on the FILLATTRS= color.

When a data skin is applied, all polygon outlines are set by the skin
and the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the polygon fill, outline, and label, when
these attributes are displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip You can use the FILLATTRS= option to set transparency for just the filled
polygon areas. You can combine this option with FILLATTRS= to set one
transparency for the polygon outline and label and a different transparency
for the polygon fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
POLYGONPLOT Statement 635

DISCRETEOFFSET=number
specifies the amount by which to offset all polygon vertices from discrete X values,
from discrete Y values, or from both.

Default 0 (all polygon vertices are centered on the discrete X values, on the
discrete Y values, or on both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right on discrete X values and up on
discrete Y values. If option REVERSE=TRUE is specified in the
layouts axis options, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Interaction This option is ignored when the XOFFSET= or YOFFSET= option is

specified.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets in order to accommodate the discrete
offset.

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display the polygon outline, fill, or both.
STANDARD
displays the polygon outline
ALL
displays the polygon outline and fill
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays the polygon outline

FILL displays the polygon fill

Default STANDARD

Tip Use the DATASKIN= , OUTLINEATTRS= , and FILLATTRS= options to

control the appearance of the fill and outline.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled polygon areas.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

For grouped data, the Color attribute of GraphData1–GraphDataN

style elements.

Interactions For this option to have any effect, the fill must be enabled by the ODS
style or by the DISPLAY= option.
636 Chapter 6 • Plot Statements

When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is
ignored, and the polygon fill colors vary according to the gradient.

Tip The DATATRANSPARENCY= option sets the transparency for the

polygon fill and outline. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines but
a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Fill Options” on page 1348 for available fill-options.

GROUP=column | discrete-attr-var | expression

creates a separate and visually distinctive polygon for each unique grouping value.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default The polygon attributes for each unique group value are derived from
the GraphData1–GraphDataN and GraphMissing style elements. If
the polygon is filled, then the COLOR attribute is used for the
polygon fill and the CONTRASTCOLOR attribute is used for the
polygon outline. If the polygon is not filled, then the
CONTRASTCOLOR and PATTERN attributes are used for the
polygon outline.

Requirement The group value must remain constant for the same ID value.
Otherwise, the results are unpredictable.

Interactions If a discrete attribute map variable is specified, then the colors and
outline patterns are mapped according to the associated
DISCRETEATTRMAP statement. See “DISCRETEATTRMAP
Statement” on page 1287. Otherwise, the colors and outline patterns
are mapped according to data order.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Note The group values should remain constant for the same ID value.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Restriction This option is ignored and missing group values are not displayed when
ID= and GROUP= specify the same values.
POLYGONPLOT Statement 637

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

LABEL=column | expression
specifies the label for the polygon.

Default No label is displayed for the polygon

Note The label text should be the same for all of the observations for a polygon
ID. When different labels are specified for the same ID, the label that is
specified in the first observation for that ID is used.

Tips The default label text color is based on the use of polygon fill and outline
colors, and on whether the GROUP= or COLORRESPONSE= option is
specified. To change the label text color and font, use the LABELATTRS=
option.

Use the LABELLOCATION= and LABELPOSITION= options to change

the location of the polygon label.

For long labels, use the LABELSPLIT=, LABELSPLITCHAR=,

LABELSPLITCHARTDROP=, and LABELSPLITJUSTIFY= options to
split the label into multiple lines.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the polygon label.

Defaults For non-grouped data, the GraphDataText style element, unless

COLORRESPONSE= is specified. When COLORRESPONSE= is
specified, the label color is determined by the outline color when
outlines are displayed, or by the GraphDataText style element color
attribute when outlines are not displayed. All other text attributes are
derived from the GraphDataText style element.

For grouped data, the GraphData1–GraphDataN style elements. The

label color is determined by the contrast color attribute.

Interaction If one or more text options are specified and they do not include all of
the font properties (such as color, family, size, weight, and style), then
the non-specified properties are derived from the GraphLabelText style
element or from a GraphData1–GraphDataN style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

638 Chapter 6 • Plot Statements

LABELLOCATION=INSIDEBBOX | OUTSIDEBBOX | OUTSIDE

specifies the location of the polygon label.

INSIDEBBOX locates the label inside the bounding box of the polygon.
OUTSIDEBBOX locates the label outside the bounding box of the polygon
but inside the plot area.
OUTSIDE locates the label outside the plot area.

Default INSIDEBBOX

Restriction OUTSIDE cannot be used when the POLYGONPLOT statement is

used in multi-cell layouts such as LATTICE, DATAPANEL, or
DATALATTICE in which the axes might be external to the grid.

Interaction The LABEL= option must be specified for this option to have any
effect.

Tip The label's exact position is relative to the polygon's X and Y data
ranges and is determined by the combination of this option and the
LABELPOSITION= option.

LABELPOSITION=CENTER | XMIN | XMAX | YMIN | YMAX

specifies the position of the polygon label with respect to the label location.

CENTER centers the label in the polygon’s bounding box

XMIN positions the label at the polygon's minimum X value and centers it
in the Y-value range
XMAX positions the label at the maximum X value and centers it in the Y-
value range
YMIN positions the label at the minimum Y value and centers it in the X-
value range
YMAX positions the label at the maximum Y value and centers it in the X-
value range
The following figure shows the label positions for each of the label locations that are
specified by the LABELLOCATION= option.
POLYGONPLOT Statement 639

Defaults CENTER when LABELLOCATION=INSIDEBBOX

YMAX when LABELLOCATION=OUTSIDEBBOX or

LABELLOCATION=OUTSIDE

Restriction CENTER is valid only when LABELLOCATION=INSIDEBBOX.

Interaction When LABELLOCATION=OUTSIDE, increasing label length might

cause the available plot area to decrease.

Tip When LABELLOCATION=OUTSIDE, the polygon label might collide

with the axis tick values on the orthogonal axis. In that case, if the
secondary orthogonal axis is not being used, specify the opposite end
of the axis. Otherwise, change LABELLOCATION= to INSIDEBBOX
or OUTSIDEBBOX.

LABELSPLIT=TRUE | FALSE
specifies whether to split the polygon label at the specified split characters. When
this option is set to TRUE, the polygon label is split unconditionally at each
occurrence of any of the specified split characters.

Default FALSE. The polygon label is not split.

Requirement The LABEL= option must also be specified.

Interaction The LABELSPLITCHAR= option specifies one or more characters

on which splits can occur.

See “boolean ” on page 1339 for other Boolean values that you can use.
640 Chapter 6 • Plot Statements

LABELSPLITCHAR="character-list"
specifies one or more characters on which the polygon label can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
polygon label. In that case, all of the specified split characters together are treated as
a single split character.
When LABEL= is specified and LABELSPLIT=TRUE, the polygon label is split
unconditionally at each occurrence of any of the specified split characters. If the
polygon label does not contain any of the specified characters, then the label is not
split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
labelsplitchar="abc"

The LABEL= option and the LABELSPLIT=TRUE option must

also be specified.

Interaction The LABELSPLITCHARDROP= option specifies whether the split

characters are included in the polygon label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the LABELSPLITJUSTIFY= option to specify the justification

of the strings in the polygon label block.

LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the polygon label.
TRUE
drops the split characters from the polygon label.
FALSE
includes the split characters in the polygon label. When LABELSPLIT=TRUE
and LABELSPLITCHARDROP=FALSE, each split character remains as the last
character in the current line. The characters that follow the split character, up to
and including the next split character, are then wrapped to the next line.
The following figure shows an example of polygon label Product*Group*A split on
* when LABELSPLITCHARDROP=TRUE and
LABELSPLITCHARDROP=FALSE.
POLYGONPLOT Statement 641

Default TRUE. The split characters are dropped from the polygon label.

Requirement The LABEL= option and the LABELSPLIT=TRUE option must also
be specified.

Interaction The LABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

LABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the polygon label blocks.
AUTO
justifies the labels based on the LABELPOSITION= option, as shown in the
following table.

LABELPOSITION= Value Justification

XMIN RIGHT

XMAX LEFT

CENTER, YMAX, or YMIN CENTER

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which LABELPOSITION=YMAX.
Note: The gray vertical line at the bottom of each label represents the horizontal
center of the text box for reference.

In this case, because LABELPOSITION=YMAX, AUTO centers the lines of text.

The text box is anchored the same way that the unsplit text is anchored.

Default AUTO

Requirement The LABEL= option and the LABELSPLIT=TRUE option must also
be specified.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
642 Chapter 6 • Plot Statements

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the polygon outline.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, unfilled polygons use both the CONTRASTCOLOR

and PATTERN attributes of the GraphData1–GraphDataN style
elements. Filled polygons use only the CONTRASTCOLOR attribute.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or by the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the polygon outline colors vary according to the gradient.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Line Options” on page 1349 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.
POLYGONPLOT Statement 643

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either the
ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles ID, COLORRESPONSE, GROUP, LABEL, and
URL.

ROTATE=numeric-column | numeric-constant | expression

specifies the angle of rotation for the polygon, measured in degrees. Positive angles
rotate the image counter clockwise, and negative angles rotate the image clockwise.
The angle specification can exceed 360 degrees in absolute value.

Default 0. No rotation is performed.

Interaction When this option is specified, the LABELLOCATION= and

LABELPOSITION= options are overridden with
LABELLOCATION=INSIDEBOX and LABELPOSITION=CENTER.

Notes The rotation angle is measured in screen coordinates.

A missing value in the rotation data is treated as 0.

Rotating a polygon does not change data ranges that are reported to the
axes. As a result, clipping might occur in some cases.

ROTATELABEL=AUTO | NONE | VERTICAL

specifies the rotation of the polygon label with respect to the rotation of the polygon.
AUTO
rotates the label with the rotation of the polygon.
644 Chapter 6 • Plot Statements

NONE
does not rotate the label with the rotation of the polygon. The label position
remains fixed regardless of the polygon rotation.
VERTICAL
rotates the label to a vertical position.

Restriction VERTICAL is valid only when the polygon is not rotated

(ROTATE=0). When the polygon is rotated,
ROTATELABEL=VERTICAL is ignored, and the default (AUTO)
is used instead.

Default AUTO

Interaction The LABEL= option must be specified for this option to have any
effect.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the polygon.
If you use this option, it replaces all of the information that is displayed by default.
You can specify roles for columns that do not contribute to the polygon plot along
with roles that do.
(role-list)
an ordered, space-separated list of unique POLYGONPLOT and user-defined
roles. POLYGONPLOT roles include ID, COLORRESPONSE, GROUP,
LABEL, and URL.

Tip User-defined roles are defined with the ROLENAME= option.

Example The following example displays the columns that are assigned to the
roles ID and URL, and the columns XOffset and YOffset in the data
tips. The XOffset and YOffset columns are not assigned to any
predefined POLYGONPLOT role, so they must first be assigned a
role:

ROLENAME=(TIP1=XOFFSET TIP2=YOFFSET)
TIP=(ID TIP1 TIP2 URL)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns that are assigned to the following roles are
automatically included in the data tip information: ID,
COLORRESPONSE or GROUP, LABEL, and URL.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.
POLYGONPLOT Statement 645

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

URL=string-column
specifies an HTML page that is displayed when the polygon is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
polygon that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate a plot with selectable polygons, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Note The URL values should remain constant for the same ID value.
646 Chapter 6 • Plot Statements

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

XOFFSET=numeric-column | expression
specifies an individual offset from the discrete X value on each vertex of the
polygon.

Default 0 (all polygon vertices are centered on the discrete X values)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right on discrete X values. If option
REVERSE=TRUE is specified in the layout's X-axis options, then the
offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Interaction This option overrides the DISCRETEOFFSET= option.

Tip Setting the discrete offset for the plots does not affect the axis
minimum offset and maximum offset. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets in order to accommodate the discrete
offset.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YOFFSET=numeric-column | expression
specifies an individual offset from the discrete Y value on each vertex of the
polygon.

Default 0 (all polygon vertices are centered on the discrete Y values)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. A positive offset is up on discrete Y values. If option
REVERSE=TRUE is specified in the layout's Y-axis options, then the
offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Interaction This option overrides the DISCRETEOFFSET= option.

POLYGONPLOT Statement 647

Tip Setting the discrete offset for the plots does not affect the axis
minimum offset and maximum offset. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets in order to accommodate the discrete
offset.

Details

Overview
The POLYGONPLOT statement draws a polygon from a series of X-Y value pairs that
are stored in a SAS data set. The first X-Y value pair defines the starting point of the
polygon. The next X-Y pair in the data specifies the coordinates of the first vertex. A
line segment is drawn from the starting point to the first vertex. For subsequent X-Y
pairs, a line segment is drawn from the previous vertex to the current vertex. This pattern
repeats until all of the segments have been drawn. If the last segment does not close the
polygon, then the POLYGONPLOT statement automatically draws a segment from the
last vertex back to the starting point in order to close the polygon.
Using the POLYGONPLOT statement, you can draw any data-driven shape on your
graph, which enables you to highlight data features, outline data boundaries, and so on.
Unlike DRAW statements in a BEGINPLOYGON block, the following conditions apply
to the POLYGONPLOT statement:
• you need to modify only the polygon data in the graph data set to modify the
polygon. You do not need to modify the template code. (See “BEGINPOLYGON
Statement” on page 1199.)
• you can draw polygons between plots. The BEGINPOLYGON block and DRAW
statements can draw polygons only on top of or behind the graph.

Requirements for the Polygon Data Set

In the simplest case of a single polygon, your data set must provide an X, Y, and ID
column that stores the X-Y values and the ID for your polygon. The X-Y values in the
first data-set observation must specify the starting point of your polygon. The X-Y
values in the subsequent observations must provide the coordinates of each vertex in the
order in which the polygon is to be drawn. There should be no gaps in the data. If your
last observation does not close the polygon, then the POLYGONPLOT statement
automatically draws a segment from your last vertex back to the starting point in order to
close the polygon.
If you want to draw multiple polygons, then your ID column must specify a unique
identifier value for each polygon. The identifier value associates the observations in the
data set with a specific polygon. All of the observations for each individual polygon
must be grouped together by ID and must be arranged in the order in which the polygon
segments are to be drawn.
The following table shows additional columns that you can use to customize your
polygon plot.

Column* Type Description

COLORRESPONSE Numeric Specifies the numeric column or range

(p. 633) attribute map variable that is used to
determine the polygon colors.
648 Chapter 6 • Plot Statements

Column* Type Description

GROUP (p. 636) Numeric or character Creates a separate polygon color, outline
pattern, or both for each unique grouping that
is specified.

LABEL (p. 637) Numeric or character Specifies the label for the polygon.

URL (p. 645) Character Specifies an HTML page that is displayed

when the polygon is selected.

XOFFSET (p. 646) Numeric Specifies an individual offset from the discrete
X value on each vertex of the polygon.

YOFFSET (p. 646) Numeric Specifies an individual offset from the discrete
Y value on each vertex of the polygon.

* You can specify any valid column name for these columns in your data set.

Drawing a Single Polygon

For a single polygon, the polygon data set contains an X column and Y column that
define the polygon vertices, and an ID column that specifies a constant value. The
polygon segments are drawn in the order in which they occur in the data. If the polygon
overlaps any graphics elements that were drawn earlier, those elements are obscured. In
that case, you can use transparency to enable the underlying graphics elements to show
through.
Here is example data for a simple four-sided polygon that is identified as P1 and that
starts at point X=40, Y=100.
Polygon Data

Obs id x y

1 P1 40 100
2 P1 20 220
3 P1 160 200
4 P1 180 80
5 P1 40 100
POLYGONPLOT Statement 649

The following figure shows how the polygon is drawn. Grid lines are provided to help
you locate the polygon vertices in the output.

The polygon starting point is X=40, Y=100 (shown in red). From the starting point, the
segments are drawn in data order. Data order is in a clockwise direction, as indicated by
the gray arrow. Although the last observation (X=40, Y=100) is provided in this
example, it is not required. If the last observation is not provided in the data, then the
POLYGONPLOT statement draws the last segment automatically in order to close the
polygon.
For an example, see “Example: Drawing a Simple Polygon That Highlights Data” on
page 652.

Drawing Multiple Polygons

For multiple polygons, the POLYGONPLOT data ID column specifies a unique
identifier for all of the observations that are associated with each polygon. The X and Y
columns specify the polygon vertices. The polygons are overlaid on the graph in the
order in which they occur in the data. For overlapping polygons, each polygon obscures
part or all of the polygons and graphics elements that were drawn before it. In that case,
you can use transparency to enable the underlying polygons and graphics elements to
show through.
Here is example data for three separate polygons.
Polygon Data

Obs id x y label

1 1 0 0 ID=1
2 1 20 0
3 1 20 30
4 1 0 30
5 1 0 0
6 2 30 0 ID=2
7 2 50 0
8 2 40 30
9 2 30 0
650 Chapter 6 • Plot Statements

10 3 60 0 ID=3
11 3 80 5
12 3 80 15
13 3 70 30
14 3 60 30
15 3 60 0

In addition to the ID column, X column, and Y column, the Label column is added to
label the polygons in the output. Notice that the observations for each ID value are
grouped together in the data set. The observations for each ID must occur contiguously
in the data. Otherwise, unexpected results might occur.
The following figure shows how the polygons are drawn from this data. Grid lines are
provided to help you locate the polygon vertices in the output.

The polygons are drawn in the order in which they appear in the data: rectangle (ID=1),
triangle (ID=2), and polygon (ID=3). The red dot on each shape indicates the starting
point for that shape. The gray arrow indicates the direction in which the segments are
drawn for each shape.

Drawing Holes in a Polygon

Using the POLYGONPLOT statement, you can draw one or more holes inside a
polygon. To create data for a polygon with one or more holes:
1. Specify the X and Y values for the outer polygon.
2. To start the data for a hole, add an observation that has missing X and Y values. The
missing X and Y values signal the POLYGONPLOT statement that the observations
that follow define the data for a hole.
3. Specify the X and Y values for the hole polygon.
4. Repeat Steps 2 and 3 for each additional hole.

Here is example data for a simple polygon that has two holes.
Polygon Data

Obs id x y
POLYGONPLOT Statement 651

1 1 1 1
2 1 9 1
3 1 7 5
4 1 9 9
5 1 1 9
6 1 3 5
7 1 1 1
8 1 . .
9 1 3 8
10 1 7 8
11 1 6 6
12 1 4 6
13 1 3 8
14 1 . .
15 1 3 2
16 1 7 2
17 1 6 4
18 1 4 4
19 1 3 2

Observations 1–7 specify the data for the outer polygon. In observation 8, the X and Y
values are missing, which indicates the start of the data for the first hole. Observations
9–13 define the data for the first hole polygon. Observation 14 indicates the start of the
data for the second hole, which is defined by observations 15–19.
The following figure shows how the polygon is drawn from this data. Grid lines are
provided to help you locate the polygon vertices in the output.

The outer polygon is drawn first, starting at point X=1, Y=1. The segments are drawn in
data order, which is in a counterclockwise direction as indicated by the gray arrow. The
first hole is drawn next, starting at point X=3, Y=8. Its segments are drawn in a
clockwise direction. The second hole is drawn last, starting at point X=3, Y=2. Its
segments are drawn in a counterclockwise direction.
652 Chapter 6 • Plot Statements

Example: Drawing a Simple Polygon That

Highlights Data

This example shows you how to use the POLYGONPLOT statement to draw a filled
polygon that highlights data in an iris petal dimension scatter plot. The polygon
surrounds the markers for the Setosa species in order to highlight the data for that
species. This example is a modified version of the example in “Example:
BEGINPOLYGON Statement” on page 1205. This version uses the POLYGONPLOT
statement instead of a BEGINPOLYGON block so that you can compare the two
methods. The following figure shows the output for this example.

Here is the SAS code for this example.

/* Generate the data for the polygon */
data polydata;
input polyID polyX polyY label $8-40;
datalines;
1 9 2 Setosa
1 13 5
1 16 7
1 17 6
1 20 5
1 20 1
1 17 1
1 15 0
1 14 0
1 11 0
;
run;
Example: Drawing a Simple Polygon That Highlights Data 653

/* Concatenate the SASHELP.IRIS and polygon data into

data set IRIS */
data iris;
set sashelp.iris polydata;
run;

/* Create the template for the graph */

proc template;
define statgraph discretelegend;
begingraph / drawspace=datavalue;
entrytitle "Iris Petal Dimensions";
layout overlayequated / equatetype=equate;
scatterplot x=petallength y=petalwidth / name="s"
group=species includemissinggroup=false;
ellipse x=petallength y=petalwidth / type=predicted alpha=.2
name="p80" legendlabel="80%" outlineattrs=graphconfidence;
ellipse x=petallength y=petalwidth / type=predicted alpha=.05
name="p95" legendlabel="95%" outlineattrs=graphconfidence2;
polygonplot x=polyX y=polyY id=polyID / display=(fill)
fillattrs=(color=yellow transparency=0.75)
label=label labellocation=outsidebbox labelposition=ymax;
discretelegend "s" / title="Species:";
discretelegend "p80" "p95" /across=1 autoalign=(topleft)
location=inside;
endlayout;
entryfootnote halign=left "Fisher's Iris Data" ;
endgraph;
end;
run;

/* Generate the graph */

proc sgrender data=iris template=discretelegend;
run;

Details
To draw a single polygon, the data set must provide an X column, a Y column, and an ID
column. The data specifies the polygon vertices around the Setosa data in a clockwise
direction. A Label column is added to provide a label for the polygon in the plot output.
Concatenation of the Polydata and Sashelp.Iris data sets results in missing values for the
SCATTERPLOT statement grouping variable in the Iris data set. By default, the
SCATTERPLOT statement includes missing group values. To exclude the missing group
values, the INCLUDEMISSINGGROUP=FALSE option is added to the
SCATTERPLOT statement.
In the POLYGONPLOT statement, the DISPLAY= option specifies the polygon fill only.
The FILLATTRS= option specifies the fill color as yellow and a fill transparency of
0.75. Rather than using draw statements to draw an annotation for the polygon, this
example uses the POLYGONPLOT statement label feature to label the polygon. The
LABEL= option specifies the column in the data set that contains the polygon label text.
The LABELLOCATION= and LABELPOSITION= options place the polygon label
outside of and above the polygon’s bounding box.
To draw the polygon, the POLYGONPLOT statements starts at X=9, Y=2, and draws a
segment between each vertex in data order. The last vertex, X=11, Y=0, does not close
654 Chapter 6 • Plot Statements

the polygon. To close the polygon, the POLYGONPLOT statement draws a segment
between X=11, Y=0 and X=9, Y=2 automatically.
If you want to highlight the Versicolor data instead of the Setosa data, then you need
only modify the data in the Polydata data set to draw a polygon around the Versicolor
data instead. You do not have to make any changes to the template code.

REFERENCELINE Statement
Creates a horizontal or vertical reference line.
Requirement: A REFERENCELINE statement can be used within a 2-D layout (OVERLAY,
OVERLAYEQUATED, DATALATTICE, or DATAPANEL) only.
Note: Specifying the X= option creates a line perpendicular to the X-axis at an X-intercept.
Specifying the Y= option creates a line perpendicular to the Y-axis at a Y-intercept.

Syntax
REFERENCELINE X=x-axis-value | column | expression </option(s)>;
REFERENCELINE Y=y-axis-value | column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the reference line data is to be considered when
determining the data range for the axis.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the reference lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the reference line.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the reference line.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
CURVELABEL="string" | column | expression
specifies a label for the reference line or lines.
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the reference line label(s).
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the reference line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN
specifies the position of the reference line label relative to the reference line.
REFERENCELINE Statement 655

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the reference line label at the specified split
characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the reference line label can be split
if needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the reference line label
text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the reference line label
block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all reference lines from the specified values
when the X or Y axis is discrete.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
You must use either the X= or the Y= argument in the REFERENCELINE statement.
X=x-axis-value | column | expression
specifies the X intercept of the reference line or lines.

Requirements If X is not specified, then Y must be specified.

Values must be the same type as the data type of the X axis. For
example, you should use numeric SAS date or time values (or SAS
date/time constants) for a time axis.

Unformatted numeric values do not map to a formatted discrete axis.

When the X axis is a discrete axis, the X axis value must be the
formatted value that appears on the X axis. If a column is specified
for the values in that case, then the specified column must use the
same format that is used for the X axis.

Note When a character value is specified, leading blanks are honored and
trailing blanks are ignored when the specified value is compared
with the data values.

Tips By default, if the value that is specified for the X= argument is

outside of the data range, then the data range is extended to include
the specified intercept. You can change this behavior with the
CLIP= option.

You can use the COLN() or COLC() function in an EVAL()

expression to specify multiple reference lines on the X axis. See
“GTL Functions Used with the EVAL Function” on page 1324.
656 Chapter 6 • Plot Statements

Y=y-axis-value | column | expression

specifies the Y intercept of the reference line or lines.

Requirements If Y is not specified, then X must be specified.

Values must be the same type as the data type of the Y axis.

Unformatted numeric values do not map to a formatted discrete axis.

When the Y axis is a discrete axis, Y axis value must be the
formatted value that appears on the Y axis. If a column is specified
for the values in that case, then the specified column must use the
same format that is used for the Y axis.

Note When a character value is specified, leading blanks are honored and
trailing blanks are ignored when the specified value is compared
with the data values.

Tips By default, if the value that is specified for the Y= argument is

outside of the data range, then the data range is extended to include
the specified intercept. You can change this behavior with the
CLIP= option.

You can use the COLN() or COLC() function in an EVAL()

expression to specify multiple reference lines on the Y axis. See
“GTL Functions Used with the EVAL Function” on page 1324.

Optional Arguments
CLIP=TRUE | FALSE
specifies whether the reference line data is to be considered when determining the
data range for the axis.
FALSE
specifies that the reference line values are to be considered when the axis range is
determined. The reference lines are drawn as follows based on the axis type:
• For a discrete axis, the reference line values that are not already on the axis
are added to the end of the axis data list. When applicable, the axis values are
then sorted:
• If the axis values are numeric values, then they are sorted ordinally.
• If the axis values are character values and a sorting option is applied to the
axis, then they are sorted as specified by the sorting option.
Reference lines are then drawn at the specified locations.
• For a linear, log, or time axis, a new axis data list is created by performing a
mathematical union of the data values and the reference line values. The
reference lines are then drawn at the locations specified.
TRUE
specifies that the reference line values are not to be considered when the axis
range is determined. The reference lines are drawn as follows based on the axis
type:
• For a discrete axis, if the reference line value exactly matches a value on the
axis, a reference line is drawn at that location. Otherwise, the reference line is
not drawn.
REFERENCELINE Statement 657

Note: If the axis values are formatted, then the reference line value must
exactly match the formatted axis value in order for the line to be drawn.
• For a linear, log, or time axis, if the reference line value is within the axis
data range, then the reference line is drawn at the specified location.
Otherwise, the reference line is not drawn.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABEL="string" | column | expression

specifies a label for the reference line or lines.

Requirement If you use the COLN() or COLC() function in the X= or Y= option to

specify multiple reference line intercepts, then you must use the
COLC() function in the CURVELABEL= option to specify exactly
one label for each reference line intercept. Otherwise, this option is
ignored. See “GTL Functions Used with the EVAL Function” on
page 1324.

Interactions If the X or Y argument specifies a value, then use "string".

If the X or Y argument specifies a column, then use a column to

define the label for each value.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the reference line label(s).

Default The GraphValueText style element.

Interaction For this option to take effect, the CURVELABEL= option must also be
used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the reference line label relative to the plot area.

INSIDE locates the labels inside the plot area

OUTSIDE locates the labels outside the plot area

Default OUTSIDE

Restriction OUTSIDE cannot be used when the REFERENCELINE is used in

multicell layouts such as LATTICE, DATAPANEL, or
DATALATTICE, where axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.
658 Chapter 6 • Plot Statements

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the line labels
appear.

See “Location and Position of Curve Labels” on page 185

CURVELABELPOSITION=AUTO | MAX | MIN

specifies the position of the reference line label relative to the reference line.
AUTO
automatically positions the line label near the line boundary along unused axes
whenever possible (typically Y2 and X2) in order to avoid collision with tick
values.

Restriction This option is used only when CURVELABELLOCATION=

OUTSIDE.

MAX
forces the line label to appear near maximum line values (typically, the top or
right).
MIN
forces the line label to appear near minimum line values (typically, the bottom or
left).

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

MAX when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION=

INSIDE is specified.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the line label
appears.

Note When you specify TICKVALUELIST=, VIEWMAX=, or

VIEWMIN= in an axis statement, the data points that are used to
determine the position of the reference line label might fall outside of
the graph area. In that case, the reference line label might not be
displayed or might be positioned incorrectly.

See “Location and Position of Curve Labels” on page 185

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the reference line label at the specified split characters.
When a reference line label is split, the label is split on each occurrence of the
specified split characters.

Default FALSE. The reference line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.
REFERENCELINE Statement 659

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the reference line label can be split if
needed. When multiple split characters are specified, each character in the list is
treated as a separate split character unless the specified characters appear
consecutively in the reference line label. In that case, all of the specified split
characters together are treated as a single split character.
When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the
reference line label is split unconditionally at each occurrence of any of the specified
split characters. If the reference line label does not contain any of the specified
characters, then the label is not split.
"character-list"
a list of one or more characters with no spaces between them enclosed in
quotation marks.

Example To specify the split characters a, b, and c:

curvelabelsplitchar="abc"

Default A blank space

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the reference line label or are
dropped.

Notes When multiple characters are specified, the order of the characters in
the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the reference line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the reference line label text.
TRUE
drops the split characters from the reference line label text.
FALSE
includes the split characters in the reference line label text. When
CURVELABELSPLIT=TRUE and
CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split character, up
to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a Y-axis reference line label with the
following specifications:
• CURVELABELPOSITION=MAX
660 Chapter 6 • Plot Statements

• CURVELABEL="Product*Group*A"
• CURVELABELSPLIT=TRUE
• CURVELABELSPLITCHARDROP=TRUE | FALSE
• CURVELABELSPLITCHAR="*"
Note: The horizontal line to the left of the label represents the maximum end of the
reference line for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the reference line label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the reference line label block.
AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown in
the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which CURVELABELPOSITION=MAX.
Note: The horizontal line to the left of each label represents the maximum end of the
reference line for reference.
REFERENCELINE Statement 661

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the reference lines. The following figure shows a
blue reference line with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the GraphSkins:DataSkin
style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the LINEATTRS= color.

DATATRANSPARENCY=number
specifies the degree of the transparency of the reference line.

Default 0
662 Chapter 6 • Plot Statements

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the reference line label.

DISCRETEOFFSET=number
specifies an amount to offset all reference lines from the specified values when the X
or Y axis is discrete.

Default 0 (no offset, all reference lines are centered on discrete X or Y values)

Range -0.5 to +0.5 where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right for a vertical reference line and up
for a horizontal reference line. If the layout's axis options set
REVERSE=TRUE, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the reference line.

Default The GraphReference style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element

.“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.
Example 1: Specifying a Single Reference Line 663

Default X

Restriction Another plot that establishes a data range for the designed axis must
be included.

Interactions This option is ignored if the X= argument is not specified.

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Restriction Another plot that establishes a data range for the designed axis must
be included.

Interactions This option is ignored if the Y= argument is not specified.

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
Reference lines are always drawn perpendicular to the axes. They are drawn from one
axis boundary to the companion boundary (X to X2 or Y to Y2). Axis offsets do not
apply to reference lines.
A REFERENCELINE statement can be used only within 2-D overlay-type layouts
(OVERLAY, OVERLAYEQUATED, or PROTOTYPE). A stand-alone plot statement
that provides a sufficient data range for determining axis extents must be included in the
layout. For example, a REFERENCELINE statement can be used with a scatter plot or a
histogram.
If a column is used to generate multiple reference lines, then the column type (numeric
or string) must agree with the type of data presented on the axis.

Examples

Example 1: Specifying a Single Reference Line

This example shows you how to draw a reference line using the REFERENCELINE
statement. The follow figure shows the output for this example.
664 Chapter 6 • Plot Statements

Example Program
Here is the SAS code for this example.
/* Create the template for the graph */
proc template;
define statgraph referenceline;
begingraph;
entrytitle "Line of Symmetry";
layout overlay / yaxisopts=(linearopts=(viewmin=0));
seriesplot x=x y=y;
referenceline x=3 /
lineattrs=(color=blue) curvelabel="X=3";
endlayout;
endgraph;
end;
run;

/* Generate the plot data */

data test;
do X=0 to 8 by 0.25;
Y=(x-3)*(x-3) + 5;
output;
end;
run;

/* Generate the graph */

proc sgrender data=test template=referenceline;
run;
Example 2: Specifying Reference Lines Using Data Columns 665

Example 2: Specifying Reference Lines Using Data Columns

This example shows you how to specify the reference-line intercept values for multiple
reference lines using columns in the data set for the graph. The following figure shows
the output for this example.

Output 6.1 Multiple Reference Lines in a Graph

Example Program
Here is the SAS code for this example.
/* Define the template for the graph */
proc template;
define statgraph referenceline;
begingraph;
entrytitle "Line of Symmetry";
layout overlay / yaxisopts=(linearopts=(viewmin=0));
seriesplot x=x y=y;
referenceline y=yR / curvelabel=label
lineattrs=(color=gray pattern=dot);
endlayout;
endgraph;
end;
run;

/* Generate the graph data */

data graphdata;
/* Plot data */
do X=0 to 8 by 0.25;
Y=(x-3)*(x-3) + 5;
output;
end;
666 Chapter 6 • Plot Statements

/* Reference line data */

x=.; y=.; yR=14; label='Y=14'; output;
x=.; y=.; yR=5; label='Y=5'; output;
run;

/* Generate the graph */

proc sgrender data=graphdata template=referenceline;
run;

Example 3: Specifying Reference Lines Using the COLN and COLC

Functions
This example shows you how to use the COLN() and COLC() functions to specify
multiple intercept values directly in the REFERENCELINE statement. This approach is
an alternative to including the reference line data in the plot data.The output is shown in
“Example 2: Specifying Reference Lines Using Data Columns” on page 665.

Example Program
Here is the SAS code for this example.
/* Create the template for the graph */
proc template;
define statgraph referencelines;
begingraph;
entrytitle "Line of Symmetry";
layout overlay / yaxisopts=(linearopts=(viewmin=0));
seriesplot x=x y=y;
/* Use COLN() to specify the intercept values */
referenceline y=eval(coln(14, 5)) /
/* Use COLC() to specify a label for each reference line */
curvelabel=eval(colc("Y=14", "Y=5"))
lineattrs=(color=gray pattern=dot);
endlayout;
endgraph;
end;
run;

/* Generate the plot data */

data test;
do X=0 to 8 by 0.25;
Y=(x-3)*(x-3) + 5;
output;
end;
run;

/* Generate the graph */

proc sgrender data=test template=referencelines;
run;

REGRESSIONPLOT Statement
Creates a fitted regression line or curve computed from input data.
Restriction: The REGRESSIONPLOT statement supports only models of one independent and
one dependent variable.
REGRESSIONPLOT Statement 667

Tip: Starting with the third maintenance release of SAS 9.4, subpixel rendering is enabled
by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page
33.For information about the ODS GRAPHICS statement SUBPIXEL= option, see
“ODS GRAPHICS Statement” in SAS ODS Graphics: Procedures Guide.

Syntax
REGRESSIONPLOT X=numeric-column | expression
Y=numeric-column | expression </<regression-options> <option(s)>>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the regression line and line label.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the regression line.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string"
specifies a label for the regression line.
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the regression line labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the regression line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the regression line label relative to the regression
line.
CURVELABELSPLIT=TRUE | FALSE
668 Chapter 6 • Plot Statements

specifies whether to split the regression line label at the specified split
characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the regression line label can be
split if needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the regression line label
text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the regression line label
block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of regression lines from just the observations that
correspond to each unique group value of the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the column for the X values.
Y=numeric-column | expression
specifies the column for the Y values.

Optional Arguments
CURVELABEL="string"
specifies a label for the regression line.

Defaults No regression line label is displayed

This option is not valid when the GROUP= option is specified.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the regression line labels.

Default The GraphValueText style element.

Interactions For this option to take effect, the CURVELABEL= option must also
be used.
REGRESSIONPLOT Statement 669

If the GROUP= option is specified, then this option is ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the regression line label relative to the plot area.
INSIDE
locates the labels inside the plot area
OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the REGRESSIONPLOT is used in

multi-cell layouts such as LATTICE, DATAPANEL, or
DATALATTICE, where axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the line labels
appear.

See “Location and Position of Curve Labels” on page 185

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the regression line label relative to the regression line.
AUTO
automatically positions the line label near the line boundary along unused axes
whenever possible (typically Y2 and X2) in order to avoid collision with tick
values.

Restriction This option is used only when CURVELABELLOCATION=

OUTSIDE.

MAX
forces the line label to appear near maximum line values (typically, upper right).
MIN
forces the line label to appear near minimum line values (typically, lower left).
START
forces the line label to appear near the beginning of the regression line.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the regression line has a
spiral shape.

END
forces the line label to appear near the end of the regression line.
670 Chapter 6 • Plot Statements

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the regression line has a
spiral shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION=INSIDE

is specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the line label
appears.

Note When you specify TICKVALUELIST=, VIEWMAX=, or

VIEWMIN= in an axis statement, the data points that are used to
determine the position of the label might fall outside of the graph area.
In that case, the regression-line label might not be displayed or might
be positioned incorrectly.

See “Location and Position of Curve Labels” on page 185

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the regression line label at the specified split characters.
When a regression line label is split, the label is split on each occurrence of the
specified split characters.

Default FALSE. The regression line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the regression line label can be split if
needed. When multiple split characters are specified, each character in the list is
treated as a separate split character unless the specified characters appear
consecutively in the regression line label. In that case, all of the specified split
characters together are treated as a single split character.
When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the
regression line label is split unconditionally at each occurrence of any of the
specified split characters. If the regression line label does not contain any of the
specified characters, then the label is not split.
REGRESSIONPLOT Statement 671

"character-list"
one or more characters with no delimiter between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the following
option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the regression line label or are
dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the regression line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the regression line label text.
TRUE
drops the split characters from the regression line label text.
FALSE
includes the split characters in the regression line label text. When
CURVELABELSPLIT=TRUE and
CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split character, up
to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a regression line label with the following
specifications:
• CURVELABELPOSITION=MAX
• CURVELABEL="Product*Group*A"
• CURVELABELSPLIT=TRUE
• CURVELABELSPLITCHARDROP=TRUE | FALSE
• CURVELABELSPLITCHAR="*"
Note: The horizontal line to the left of the label represents the maximum end of the
regression line for reference.
672 Chapter 6 • Plot Statements

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the regression line
label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the regression line label block.
AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown in
the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which CURVELABELPOSITION=MAX.
Note: The horizontal line to the left of each label represents the maximum end of the
regression line for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO
REGRESSIONPLOT Statement 673

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the regression line and line label.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

GROUP=column | discrete-attr-var | expression

creates a distinct set of regression lines from just the observations that correspond to
each unique group value of the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of line color and line pattern. Line colors vary
according to the ContrastColor attribute of the GraphData1–
GraphDataN and GraphMissing style elements. Line patterns vary
according to the LineStyle attribute of the GraphData1–GraphDataN
style elements.

Restriction The input data must be sorted by the GROUP= column.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line colors and
line patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Tip The LINEATTRS= option can be used to override the representations

that are used to identify the groups. For example,
LINEATTRS=(PATTERN=SOLID) can be used to assign the same
pattern to all of the lines, letting the line color distinguish group
values. Likewise, LINEATTRS=(COLOR=BLACK) can be used to
assign the same color to all of the lines, letting the line pattern
distinguish group values.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.
674 Chapter 6 • Plot Statements

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the regression line.

Default The GraphFit style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.
REGRESSIONPLOT Statement 675

“Line Options” on page 1349 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles are
automatically included in the data tip information.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
676 Chapter 6 • Plot Statements

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Curve")

Default The column label or column name of the column assigned to the role.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles are
automatically included in the data tip information.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Regression Options
ALPHA=positive-number
specifies the confidence level to compute.

Default 0.05

Range 0 < positive-number < 1

Tip ALPHA=0.05 represents a 95% confidence level.

CLI="name"
produces confidence limits for individual predicted values for each observation. The
confidence level is set by the ALPHA= option.

Interaction name is a unique name within the template that is case sensitive and
cannot contain spaces. It must be assigned in order for the confidence
limits to be computed. To display confidence limits, you must use this
name as the required argument of a MODELBAND statement. See the
example in “Example: REGRESSIONPLOT Statement” on page
678 .

CLM="name"
produces confidence limits for a mean predicted value for each observation. The
confidence level is set by the ALPHA= option.
REGRESSIONPLOT Statement 677

Interaction name is a unique name within the template that is case sensitive and
cannot contain spaces. It must be assigned in order for the confidence
limits to be computed. To display confidence limits, you must use this
name as the required argument of a MODELBAND statement. See the
example in “Example: REGRESSIONPLOT Statement” on page
678 .

DEGREE=non-negative-integer
specifies the degree of the polynomial.
The default value, DEGREE=1, produces a linear fit, DEGREE=2 produces a
quadratic fit, DEGREE=3 produces a cubic fit, and so on.
The value of the DEGREE=d option corresponds to one of the following
TRANSREG procedure specifications for the independent variable: SPLINE(X /
DEGREE=d) or PBSPLINE(X / DEGREE=d LAMBDA=0).

Default 1

Ranges 0–174 in the first maintenance release of SAS 9.4 and earlier releases.

0–10 starting with the second maintenance release of SAS 9.4.

FREQ=numeric-column
specifies a column in the input data set that represents the frequency of occurrence of
the current observation, essentially treating the data set as if each observation
appeared n times, where n is the value of the FREQ column for the observation.
Noninteger values of the FREQ column are truncated to the largest integer less than
the FREQ value. The observation is used in the analysis only if the value of the
FREQ column is greater than or equal to 1.
MAXPOINTS=positive-integer
specifies the maximum number of predicted points generated for the regression curve
as well as any confidence limits.

Default 201

WEIGHT=numeric-column
specifies a column in the input data set that contains values to be used as a priori
weights for a regression fit. If an observation’s weight is zero, negative, or missing,
then the observation is deleted from the analysis.

Interaction Starting with the second maintenance release of SAS 9.4, when the
CLI= option is used with this option, the confidence band for individual
predicted values is displayed as a high-low chart instead of a band.

Details
The REGRESSIONPLOT statement only supports models of one independent and one
dependent variable. For more information about the fitting methodology, see the
TRANSREG procedure in the SAS/STAT user’s guide.
In addition to the regression line, the REGRESSIONPLOT statement can compute
confidence levels for the fitted line. To display the confidence levels:
1. Use the CLI= or CLM= regression option(s) to declare a name for each confidence
level.
678 Chapter 6 • Plot Statements

2. Use MODELBAND statements to refer to the name(s) and draw a confidence

band(s) from this information.

Example: REGRESSIONPLOT Statement

The following graph was generated by the “Example Program” on page 678:

Example Program
proc template;
define statgraph regressionplot;
begingraph;
entrytitle "Regression Fit Plot";
layout overlay;
scatterplot x=weight y=mpg_highway /
datatransparency=0.7;
regressionplot x=weight y=mpg_highway /
name="fitline"
alpha=0.05 legendlabel="Regression Fit";
discretelegend "fitline";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=regressionplot;

run;
SCATTERPLOT Statement 679

SCATTERPLOT Statement
Creates a scatter plot of input data.

Syntax
SCATTERPLOT X=column | expression
Y=column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group
clusters as a fraction of the minimum interval between adjacent data values.
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
MARKERCOLORGRADIENT= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
starting with the second maintenance release of SAS 9.4, specifies the
column or range attribute map variable to use to determine the marker colors.
CONTRIBUTEOFFSETS=ALL | NONE | (axis-offset-list)
specifies whether space requirements for this plot contribute to the
calculation of the axis offsets.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the plot markers.
DATATRANSPARENCY=number
specifies the degree of the transparency of the markers, data labels, and error
bars, when displayed.
DISCRETEMARKERSIZE=number
specifies the size of a marker as a fraction of the tick spacing.
ERRORBARATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the error bars that are associated with the data
points.
ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
INDEX=positive-integer-column | expression
specifies indices for mapping marker attributes (color and symbol) to one of
the GraphData1–GraphDataN style elements.
LABELSTRIP=TRUE | FALSE
specifies whether leading and trailing blanks are stripped from marker
characters or data labels that have a fixed position before they are displayed
in the plot.
MARKERATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the data markers.
MARKERCHARACTER=column | expression
680 Chapter 6 • Plot Statements

specifies a column that defines strings that are to be used instead of marker
symbols.
MARKERCHARACTERATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the marker character specified on the
MARKERCHARACTER= option.
MARKERCHARACTERPOSITION=CENTER | TOP | BOTTOM | LEFT | RIGHT |
TOPLEFT | TOPRIGHT | BOTTOMLEFT | BOTTOMRIGHT
specifies the justification of the marker characters.
MARKERCOLORGRADIENT=numeric-column | range-attr-var | expression
in the first maintenance release of SAS 9.4 and earlier releases, specifies the
column or range attribute map variable that is used to determine the marker
colors.
MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.
MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.
MARKERSIZEMAX=dimension
for the first maintenance release of SAS 9.4 and for earlier releases, specifies
a drawing size for the largest marker when the marker size represents
response values.
MARKERSIZEMIN=dimension
for the first maintenance release of SAS 9.4 and for earlier releases, specifies
a drawing size for the smallest marker when the marker size represents
response values.
MARKERSIZERESPONSE=numeric-column | expression
for the first maintenance release of SAS 9.4 and for earlier releases, specifies
a column that is used to map the drawing size of the markers.
OUTLINEDMARKERCHARACTERS=TRUE | FALSE
specifies whether the characters that are used as marker symbols are outlined
in order to enhance their appearance in the graph.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.
SIZEMAX=dimension
starting with the second maintenance release of SAS 9.4, specifies a drawing
size for the largest marker when the marker size represents response values.
SIZEMIN=dimension
starting with the second maintenance release of SAS 9.4, specifies a drawing
size for the smallest marker when the marker size represents response values.
SIZERESPONSE=numeric-column | expression
starting with the second maintenance release of SAS 9.4, specifies a column
that is used to map the drawing size of the markers.
SUBPIXEL=AUTO | OFF
specifies whether subpixel rendering is used for image output when the
scatter plot is rendered.
USEDISCRETESIZE=TRUE | FALSE
specifies that the marker size should be based on fraction of the midpoint
spacing that is set by the DISCRETEMARKERSIZE= option.
XERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the X error bars. The error bars
are drawn from the markers to the endpoints.
SCATTERPLOT Statement 681

XERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the X error bars.
YERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the Y error bars.
YERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the Y error bars.

Axes options
CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
scatter points.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
specifies a column for marker labels. The label positions are adjusted to
prevent them from overlapping.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |
CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the markers.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label blocks.
682 Chapter 6 • Plot Statements

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all markers from discrete X values, or discrete Y
values, or both.
GROUP=column | discrete-attr-var | expression
creates a separate marker type for each unique group value of the specified
column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.
JITTER=NONE | AUTO
specifies whether to jitter data markers.
JITTEROPTS=(jitter-options)
specifies options for managing jittering.

ODS options
URL=string-column
specifies an HTML page to display when a point is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation
that is read.

Required Arguments
X=column | expression
specifies the column for the X values.
Y=column | expression
specifies the column for the Y values.

Optional Arguments
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
MARKERCOLORGRADIENT= option.
style-element
specifies the name of a style element. The style element should contain these
style attributes:
SCATTERPLOT Statement 683

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= or

MARKERCOLORGRADIENT= option must also be specified.

Tip The REVERSECOLORMODEL= option can be used to reverse the

start and end colors of the ramp assigned to the color model.

COLORRESPONSE=numeric-column | range-attr-var | expression

starting with the second maintenance release of SAS 9.4, specifies the column or
range attribute map variable to use to determine the marker colors.
Note: Starting with the second maintenance release of SAS 9.4, the
COLORRESPONSE= option replaces the MARKERCOLORGRADIENT=
option. The syntax and functionality are the same. The
MARKERCOLORGRADIENT= option is still honored, but the
COLORRESPONSE= option is preferred.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

Tip For each range in the attribute map, the RANGEALTCOLOR= or

RANGEALTCOLORMODEL= option in the RANGE statement
determines the marker colors.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.
You can use this option to add a second response variable to an analysis. For
example, in an analysis of weight by height, you can specify an age column by using
684 Chapter 6 • Plot Statements

the COLORRESPONSE= or MARKERCOLORGRADIENT= option so that the

change in the gradient color of the markers reflects the change in age.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column has
multiple values for a single GROUP value, unexpected results might
occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

Suboption COLOR= in the DATALABELATTRS= option overrides

this option for the data label color attribute.

This option overrides suboption COLOR= in the MARKERATTRS=

option and in the MARKERCHARACTERATTRS= option and
varies the marker color according to the color gradient or the attribute
map.

Note You can use MARKERCOLORGRADIENT= as an alternative to

COLORRESPONSE=. However, if you use
MARKERCOLORGRADIENT=, be aware that the TIP=,
TIPFORMAT=, and TIPLABEL= options recognize
MARKERCOLORGRADIENT as the color role, not
COLORRESPONSE.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

Starting with the second maintenance release of SAS 9.4, when this
option is in effect and error bars are displayed, the error bars derive
their color from the markers. To set a fixed color for the error bars,
use the ERRORBARATTRS= option.

If the MARKERCHARACTER= option is also specified, then the

gradients that would be applied to the markers are applied to the text
strings.

CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when GROUPDISPLAY=CLUSTER.
SCATTERPLOT Statement 685

AUTO uses the discrete axis for clustering groups when only one axis is
discrete. Uses the X axis for clustering if both axes are discrete or
interval.
X|Y uses the X or Y axis for clustering groups.

Default AUTO

Interaction The GROUPDISPLAY= option must be set to CLUSTER for this

option to have any effect.

CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters as a
fraction of the minimum interval between adjacent data values.

Default 0.85

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.

Interactions For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

When GROUPDISPLAY=CLUSTER and CLUSTERWIDTH= are in

effect for interval data, the size of the markers in each cluster might be
reduced to no less than 5 pixels in order to display the cluster within
the smallest effective midpoint space. If you need larger markers in
that case, use the MARKERATTRS= option to specify a larger marker
size.

CONTRIBUTEOFFSETS=ALL | NONE | (axis-offset-list)

specifies whether space requirements for this plot contribute to the calculation of the
axis offsets.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
This plot’s layout container queries each of its plots for a preferred offset and
includes all of the offsets in the axis offset calculations. If the DATALABEL= or
MARKERCHARACTER= option is specified for this plot, this plot might request a
preferred offset that prevents the clipping of any data labels or marker character
686 Chapter 6 • Plot Statements

strings that appear at the ends of the axes. The requested offset is based on the
longest string. If the label or marker character lengths vary significantly, the result is
wasted space when the shorter strings appear near the ends of the axes. In that case,
you can use the CONTRIBUTEOFFSETS= option to modify or eliminate this plot’s
contribution to the offset calculations in order to reclaim that space.
ALL
the space requirements for this plot are contributed to the axis offset calculations.
NONE
the space requirements for this plot are not contributed to the axis offset
calculations.
(axis-offset-list)
a space-delimited list of specific contributions that this plot makes to the axis
offset calculations. The list is one or more of the following values, enclosed in
parentheses:

XMAX the space requirements for this plot are contributed to the X-axis
offset calculation for the maximum end.
XMIN the space requirements for this plot are contributed to the X-axis
offset calculation for the minimum end.
YMAX the space requirements for this plot are contributed to the Y-axis
offset calculation for the maximum end.
YMIN the space requirements for this plot are contributed to the Y-axis
offset calculation for the minimum end.

Default ALL

Interaction Offsets that are set in the layout axis options are always honored,
regardless of the setting on this option.

Note This option does not affect offset requests from other plots.

DATALABEL=column | expression
specifies a column for marker labels. The label positions are adjusted to prevent
them from overlapping.

Default No data labels are displayed

SCATTERPLOT Statement 687

Interactions If a numeric column is specified and the column has no format, then a
BEST6 format is applied.

This option is ignored if the MARKERCHARACTER= option is used,

which displays labels instead of the markers.

Tip Use this option to display labels for the markers. The position of the
labels are adjusted to prevent the labels from overlapping. If you want
labels displayed instead of markers, then use the
MARKERCHARACTER= option.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the DATALABEL= option must also be
specified.

This option is ignored if the MARKERCHARACTER= option is

specified.

Note When the DATALABELPOSITION=AUTO option is in effect, in

some cases, the data label font size might be reduced in order to avoid
overlapping labels and markers.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element

“Text Options” on page 1351 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |

CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the markers.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.

Default FALSE

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
data label. In that case, all of the specified split characters together are treated as a
single split character.
688 Chapter 6 • Plot Statements

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters. If the
data label does not contain any of the specified characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

Interactions This option has no effect if DATALABELPOSITION=AUTO or if

DATALABELSPLIT=FALSE.

The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels. The split
characters are specified by the DATALABELSPLITCHAR= option.
TRUE
drops the split characters from the data label.
FALSE
includes the split characters in the data label. When DATALABELSPLIT=TRUE
and DATALABELSPLITCHARDROP=FALSE, each split character remains as
the last character in the current line. The characters that follow the split character,
up to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a data label with the following
specifications:
• the data label text for this label is Product*Group*A
• DATALABELSPLIT=TRUE
• DATALABELSPLITCHARDROP=TRUE | FALSE
• DATALABELSPLITCHAR="*"
SCATTERPLOT Statement 689

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from the

label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains as the
last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Interactions This option has no effect unless DATALABELSPLIT=TRUE.

The DATALABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.
AUTO
justifies the labels based on the DATALABELPOSITION= option, as shown in
the following table.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which DATALABELPOSITION=TOP.
Note: The gray vertical line at the bottom of each label represents the horizontal
center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of text.

The text box is anchored the same way that the unsplit text is anchored. For example,
if DATALABELPOSITION=TOP, then the bottom center of the text box is
positioned at the top of the marker.

Default AUTO
690 Chapter 6 • Plot Statements

Interaction This option has no effect if DATALABELPOSITION=AUTO or if

DATALABELSPLIT=FALSE.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the plot markers. The following figure shows large
CIRCLEFILLED markers with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

When a data skin is applied, all marker outlines are set by the skin,
and the outline settings are ignored from the ODS style or from
MARKERATTRS= option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the markers, data labels, and error bars,
when displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISCRETEMARKERSIZE=number
specifies the size of a marker as a fraction of the tick spacing.

Default 0.5

Range 0 to 1
SCATTERPLOT Statement 691

Requirement For this option to take effect, at least one of the axes must be discrete.

Interactions If both of the axes are discrete, then the marker size is a fraction of
the smaller tick spacing.

If the X axis is a numeric interval axis and the

GROUPDISPLAY=CLUSTER option is in effect, then the marker
size is a fraction of the interval between the two closest but not
identical points in the X direction.

For this option to take effect, the USEDISCRETESIZE= option must

be set to TRUE (the default is FALSE).

DISCRETEOFFSET=number
specifies an amount to offset all markers from discrete X values, or discrete Y
values, or both. This feature is useful for graphing multiple response variables side
by side on a common axis. By default within an overlay-type layout, if a
SCATTERPLOT is used with other plots with a discrete axis, then the markers are
centered on the discrete X values, or discrete Y values, or both. Depending on the
data, the markers might be superimposed over other graph data. The following code
fragment shows the default positioning when a SCATTERPLOT is used with a
BOXPLOT:
layout overlay / cycleattrs=true
xaxisopts=(type=discrete);

scatterplot x=age y=weight;

boxplot x=age y=weight;

endlayout;

To avoid superimposed plots, you can assign a different offset to each plot statement:
layout overlay / cycleattrs=true
xaxisopts=(type=discrete);
scatterplot x=age y=weight /
discreteoffset=0.2;
boxplot x=age y=weight /
discreteoffset=-0.2;
endlayout;
692 Chapter 6 • Plot Statements

Default 0 (no offset, all markers are centered on the discrete X values, or
discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right on discrete X values and up on
discrete Y values. If the layout’s axis options set REVERSE=TRUE,
then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

ERRORBARATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the error bars that are associated with the data points.

Defaults For non-grouped data, the GraphError style element.

For grouped data, the LineStyle and LineThickness attributes of the

GraphError style element and the ContrastColor attribute of the
GraphData1–GraphDataN style elements. (The LineStyle does not
apply to the "serif" parts of the error bars.)

Interaction For this option to take effect, error bars must be displayed by the
XERRORLOWER= , XERRORUPPER= , YERRORLOWER= , or
YERRORUPPER= options.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
SCATTERPLOT Statement 693

Defaults SERIF in the first maintenance release of SAS 9.4 and earlier releases.

Starting with the second maintanance release of SAS 9.4,

GraphError:CapStyle style reference. If attribute CapStyle is not defined
in the active style, then SERIF is the default value.

Tip The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline. When this option is TRUE, the fill color and outline color for
filled markers are determined in the following ways:
• If the GROUP= option is specified, then by default, the fill color is derived
from the GraphData1–GraphDataN style elements Color attribute, and the
marker outlined color is derived from the GraphData1–GraphDataN style
elements ContrastColor attribute.
• If the GROUP= option is not specified, then by default, the fill color is
derived from the GraphDataDefault style elements Color attribute, and the
marker outlined color is derived from the GraphOutline style elements
ContrastColor attribute.
• If the COLORRESPONSE= MARKERCOLORGRADIENT= option is
specified, then the marker fill is drawn by using the mapped color that is
computed from the value of the COLORRESPONSE= or
MARKERCOLORGRADIENT= option for that observation. The marker
outline is drawn by using the MARKEROUTLINEATTRS= specification.
FALSE
draws the markers using fill or an outline, but not both.

Default FALSE

Tip To specify the marker fill and outline colors for a non-grouped plot, set this
option to TRUE, and then use the MARKERFILLATTRS= and
MARKEROUTLINEATTRS= options to specify the colors.

See GROUP= on page 694

MARKERFILLATTRS= on page 703

MARKEROUTLINEATTRS= on page 703

COLORRESPONSE= on page 683

“boolean ” on page 1339 for other Boolean values that you can use.
694 Chapter 6 • Plot Statements

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that is
read.

Default All observations have a frequency count of 1.

Restriction If the value of the numeric-column is missing or is less than 1, then the
observation is not used in the analysis. If the value is not an integer,
then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation.

GROUP=column | discrete-attr-var | expression

creates a separate marker type for each unique group value of the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color and marker symbol. Markers vary
according to the ContrastColor and MarkerSymbol attributes of the
GraphData1–GraphDataN and GraphMissing style elements.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of markers and
colors.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

The marker size is set by the MARKERATTRS= option.

The MARKERCHARACTER= , COLORRESPONSE= , and

MARKERCOLORGRADIENT= options override the group settings
for the marker symbol and marker color.

The SIZERESPONSE= and MARKERSIZERESPONSE= options

override this option’s SIZE= setting.

Tip The representations that are used to identify the groups can be
overridden. For example, each distinct group value is represented by a
different marker symbol, but the
MARKERATTRS=(SYMBOL=marker) option could be used to
assign the same symbol to all of the plot’s marker symbols, letting
marker color indicate group values. Likewise,
MARKERATTRS=(COLOR=color) could be used to assign the same
color to all markers, letting marker symbol indicate group values.

See “DISCRETEATTRVAR Statement” on page 1297

SCATTERPLOT Statement 695

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.
OVERLAY
draws markers for a given group value at the exact coordinate. Depending on the
data, markers at a given coordinate might overlap.
CLUSTER
draws markers for a given group value adjacent to each other.

Default OVERLAY

Interaction For interval data, when GROUPDISPLAY=CLUSTER is in effect, the

size of the markers in each cluster might be reduced to no less than 5
pixels in order to display the cluster within the smallest effective
midpoint space. If you need larger markers in that case, use the
MARKERATTRS= option to specify a larger marker size.

Tip Use the CLUSTERWIDTH= option to control the width of the clusters
when CLUSTER is in effect.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
DESCENDING
orders the groups within a category in descending order.

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.
696 Chapter 6 • Plot Statements

Tips Use the CLUSTERWIDTH= option to control the distance between

the group markers in a cluster.

Use the INDEX= option to alter the default sequence of visual

attributes that is assigned to the groups.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping marker attributes (color and symbol) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interactions For this option to take effect, the GROUP= option must also be
specified.

The MARKERCHARACTER=, COLORRESPONSE=, and

MARKERCOLORGRADIENT= options override the group settings
for the marker symbol and marker color.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.
SCATTERPLOT Statement 697

JITTER=NONE | AUTO
specifies whether to jitter data markers.
NONE
does not offset the data markers.
AUTO
the system determines the best way to display the data markers, based on the
settings specified by the JITTEROPTS= option.
The following figures show a side-by-side comparison of a BOXPLOT and
SCATTERPLOT in which the JITTER=NONE and JITTER=AUTO options are
specified in the SCATTERPLOT statements. The first figure shows the case in which
the Y axis is linear and the X axis is discrete.

In this case, when jittering is disabled (JITTER=NONE), markers that represent the
same Y value are overlaid, which results in some markers being obscured. When
jittering is enabled (JITTER=AUTO), one-dimensional systematic jittering occurs
along the X axis. The markers that represent approximately the same Y value are
offset along the X axis from the midpoint of that value in order to make all of the
markers visible.
The next figure shows the case in which both axes are linear.
698 Chapter 6 • Plot Statements

In this case, when jittering is disabled (JITTER=NONE), markers that represent the
same X and Y bin value are overlaid as in the previous case, which results in some
markers being obscured. However, unlike the previous case, when jittering is enabled
(JITTER=AUTO), two-dimensional random jittering occurs along the X and the Y
axes. The markers are offset randomly along both the X and Y axes in order to make
all of the markers visible.

Default NONE

Restrictions In SAS 9.4, jittering is not supported when

GROUPDISPLAY=CLUSTER. Starting with the first maintenance
release of SAS 9.4, this restriction is removed.

One-dimensional systematic jittering is not supported when the

MARKERCHARACTER= or the MARKERSIZERESPONSE=
option is in effect. Random jittering is supported in those cases when
both the X and Y axes are interval axes.

Interactions When jittering is enabled, the FREQ= option is ignored.

If this option is set to AUTO and both axes are discrete, then one-
dimensional systematic jittering occurs along the X axis.

Notes Jittering changes the default axis offsets, but it does not change the
axis data range.

When jittering is requested on a discrete axis and a large amount of

data is plotted, the jittering process can become resource-intensive. In
that case, it might take longer to render the plot.

JITTEROPTS=(jitter-options)
specifies options for managing jittering. The jitter options can be one or more of the
following values, separated by a space:
AXIS=AUTO | X | Y | BOTH
specifies the axis to use for jittering the data markers.
AUTO
the system determines the axis, based on the following criteria:
• If the X axis is discrete, then one-dimensional systematic jittering is
applied along the X axis.
• If the Y axis is discrete and the X axis is interval, then one-dimensional
systematic jittering is applied along the Y axis.
• If the X and Y axes are interval, then random jittering is applied along
both the X and Y axes.
X
jittering is on the X axis. If the X axis is discrete, then 1-dimensional
systematic jittering is applied. Otherwise, 1-dimensional random jittering is
applied.

Note If both the X and Y axes are discrete, then specifying BOTH is
equivalent to specifying X.
SCATTERPLOT Statement 699

Y
jittering is applied on the Y axis. If the Y axis is discrete, then 1-dimensional
systematic jittering is applied. Otherwise, 1-dimensional random jittering is
applied.
BOTH
specifies that random jittering is applied on both the X and Y axes. This
option applies only when both the X and Y axes are interval.

Tip When the X or Y axis is discrete, specifying BOTH is equivalent to

specifying AUTO.

Default AUTO

Restriction When the MARKERCHARACTER= or the

MARKERSIZERESPONSE= option is in effect, 1-dimensional
systematic jittering is not supported. Random jittering is supported
in those cases when both the X and Y axes are interval axes.

WIDTH=positive-number
specifies the width of the jittering space as a fraction of either the midpoint
spacing or of the minimal interval width.

Defaults 0.85 for 1-dimensional systematic jittering

0.4 for random jittering on one or both axes.

Note For a discrete axis, this option has effect only when the markers
cannot be clustered side-by-side without overlapping. When WIDTH=
is set to a value that is sufficient to eliminate marker overlap in that
case, increasing the value further has no effect.

Tip The specified number can be greater than 1.

Interaction This option is ignored when JITTER=NONE.

LABELSTRIP=TRUE | FALSE
specifies whether leading and trailing blanks are stripped from marker characters or
data labels that have a fixed position before they are displayed in the plot. The
MARKERCHARACTER= option specifies the column that provides the marker
strings that are to be used in place of marker symbols.

Default FALSE

Interactions This option effects marker strings only when the

MARKERCHARACTER= option is specified.

This option effects data labels only when DATALABEL= is specified

and DATALABELPOSITION= is not AUTO.

Tip Stripping the blanks from the numeric value strings helps center each
string relative to its data point. Stripping is useful when you want to
overlay the data values near or inside the markers for a plot.

See “boolean ” on page 1339 for other Boolean values that you can use.
700 Chapter 6 • Plot Statements

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y-column label. If a label is not defined, then the Y-column name
is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

MARKERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes of

the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions If the COLORRESPONSE= or MARKERCOLORGRADIENT=

option is specified, then this option’s COLOR= suboption is ignored.

If the MARKERCHARACTER= option is specified, then this option’s

SYMBOL= and WEIGHT= suboptions are ignored.

If FILLEDOUTLINEDMARKERS=TRUE, then this option’s

COLOR= suboption is ignored. In that case, to specify the marker fill
color, use the MARKERFILLATTRS= option instead.

This option’s COLOR= suboption overrides the default behavior for

grouped data. When the COLOR= suboption is specified in that case,
all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior for

grouped data. When the SYMBOL= suboption is specified in that
case, all markers have the same symbol, and the symbol color alone
distinguishes the markers.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Marker Options” on page 1350 for available marker-options.

MARKERCHARACTER=column | expression
specifies a column that defines strings that are to be used instead of marker symbols.
SCATTERPLOT Statement 701

Each string is centered horizontally and vertically at the data point. The data point
positions are not adjusted to prevent text overlap.

Interactions This option overrides the DATALABEL= option.

If the GROUP= option is also used, then color is displayed for a

DISCRETE legend, but the character is not displayed in the legend.
For text strings, the same colors are applied to the text strings as
would have been applied to markers.

One-dimensional systematic jittering is not supported when this option

is in effect. Random jittering is supported when this option is in effect
and both the X and Y axes are interval axes. See JITTER= on page
697.

Note If a numeric column is used, then its values are converted to strings
using the format associated with the column or BEST6 if no format is
defined.

Tips Lengthy strings might be clipped by the plot border. To reduce

clipping, you can use the OFFSETMIN= and OFFSETMAX=
suboptions of the XAXISOPTS= and YAXISOPTS= options to
increase the axis offsets.

You can use the MARKERCHARACTERPOSITION= option to

change the justification of the marker character.

You can use the OUTLINEDMARKERCHARACTERS= option to

enhance the appearance of the marker characters.

You can use the LABELSTRIP= option to strip the leading and
trailing blanks from numeric value strings in order to center each
string on its data point.

MARKERCHARACTERATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the marker character specified on the
MARKERCHARACTER= option.

Defaults For non-grouped data, the GraphDataText style element.

702 Chapter 6 • Plot Statements

For grouped data, GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the MARKERCHARACTER= option

must also be used.

When the GROUP= option is also specified, each distinct group value
might be represented by a different color (depending on the ODS style
setting or the setting on the INDEX= option). The marker character
that is associated with the group is assigned the group color.

This option’s COLOR= suboption can be used to specify a single color

for all marker characters in a grouped plot, without affecting items that
have a group color, such as error bars and marker symbols.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

MARKERCHARACTERPOSITION=CENTER | TOP | BOTTOM | LEFT |

RIGHT | TOPLEFT | TOPRIGHT | BOTTOMLEFT | BOTTOMRIGHT
specifies the justification of the marker characters. The following figure shows the
effect of each of the values on the position of marker character M.

Note: The red dots represent the center in each case for reference.

Default CENTER

Interaction This option is ignored if the MARKERCHARACTER= option is not

specified.

Tip You can use the LABELSTRIP= option to strip the leading and trailing
blanks from value strings in order to properly justify each string on its
data point.

MARKERCOLORGRADIENT=numeric-column | range-attr-var | expression

in the first maintenance release of SAS 9.4 and earlier releases, specifies the column
or range attribute map variable that is used to determine the marker colors.
Note: Starting with the second maintenance release of SAS 9.4, the
MARKERCOLORGRADIENT= option is deprecated and replaced with the
COLORRESPONSE= option. The syntax and functionality are the same. The
MARKERCOLORGRADIENT= option is still honored, but the
COLORRESPONSE= option is preferred.
SCATTERPLOT Statement 703

Note Starting with the second maintenance release of SAS 9.4, if you use the
MARKERCOLORGRADIENT= option, then be aware that the TIP=,
TIPFORMAT=, and TIPLABEL= options recognize the
MARKERCOLORGRADIENT role and not the COLORRESPONSE role.

See COLORRESPONSE= on page 683

MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.

Defaults For non-grouped data, the COLOR attribute of the GraphDataDefault

style element

For grouped data, the COLOR attribute of a GraphData1–GraphDataN

style element

Restriction The TRANSPARENCY= fill option is ignored. Use the

MARKERATTRS= option to set the marker transparency.

Interactions This option is in effect only when

FILLEDOUTLINEDMARKERS=TRUE and the DISPLAY= option
enables fill display.

When the COLORRESPONSE= or MARKERCOLORGRADIENT=

option is in effect, this option’s COLOR= specification is ignored.

Note When style-element is specified, only the style element’s COLOR

attribute is used.

See “General Syntax for Attribute Options” on page 1347

“Fill Options” on page 1348

MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, the LineThickness attritube of the GraphOutlines

style element and the ContrastColor attribute of a GraphData1–
GraphDataN style element.

Restriction The line style of the marker outline is always solid.

Interaction This option is ignored when a data skin is applied by the current style
or by the DATASKIN= option. In the latter case, the outline is set by
the data skin.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR and LINETHICKNESS attributes are used.

See “General Syntax for Attribute Options” on page 1347

“Line Options” on page 1349

MARKERSIZEMAX=dimension
for the first maintenance release of SAS 9.4 and for earlier releases, specifies a
drawing size for the largest marker when the marker size represents response values.
704 Chapter 6 • Plot Statements

Note: Starting with the second maintenance release of SAS 9.4, the SIZEMAX=
option replaces the MARKERSIZEMAX= option. The syntax and functionality
are the same. The MARKERSIZEMAX= option is still honored, but the
SIZEMAX= option is preferred.

See SIZEMAX= on page 706

“dimension” on page 1340

MARKERSIZEMIN=dimension
for the first maintenance release of SAS 9.4 and for earlier releases, specifies a
drawing size for the smallest marker when the marker size represents response
values.
Note: Starting with the second maintenance release of SAS 9.4, the SIZEMIN=
option replaces the MARKERSIZEMIN= option. The syntax and functionality
are the same. The MARKERSIZEMIN= option is still honored, but the
SIZEMIN= option is preferred.

See SIZEMIN= on page 706

“dimension” on page 1340

MARKERSIZERESPONSE=numeric-column | expression
for the first maintenance release of SAS 9.4 and for earlier releases, specifies a
column that is used to map the drawing size of the markers.
Note: Starting with the second maintenance release of SAS 9.4, the
SIZERESPONSE= option replaces the MARKERSIZERESPONSE= option. The
syntax and functionality are the same. The MARKERSIZERESPONSE= option
is still honored, but the SIZERESPONSE= option is preferred.

Interaction One-dimensional systematic jittering is not supported when this option

is in effect. Random jittering is supported when this option is in effect
and both the X and Y axes are interval axes. See JITTER= on page 697.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEDMARKERCHARACTERS=TRUE | FALSE
specifies whether the characters that are used as marker symbols are outlined in order
to enhance their appearance in the graph.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
Note: This feature is deprecated starting with the second maintenance release of
SAS 9.4. The OUTLINEDMARKERCHARACTERS= option is still honored,
but the TEXTPLOT statement is now the preferred method of creating a scatter
plot using text markers.
SCATTERPLOT Statement 705

The following figure shows the marker characters M and F displayed when
OUTLINEDMARKERCHARACTERS=FALSE (default) and when
OUTLINEDMARKERCHARACTERS=TRUE.

Default FALSE

Restriction Outline marker characters are not supported in vector graphics output.
When this option is set and vector graphics output is requested, the
graph is converted into an image instead. A note indicating the
conversion is written to the SAS log. To restore the vector graphics
output in that case, remove the
OUTLINEDMARKERCHARACTERS= option from the
SCATTERPLOT statement.

Interaction The MARKERCHARACTER= option must be specified for this option

to have any effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either the
ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page 1339 for other Boolean values that you can use.
706 Chapter 6 • Plot Statements

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, DATALABEL, MARKERCHARACTER,
COLORRESPONSE or MARKERCOLORGRADIENT,
XERRORLOWER, XERRORUPPER, YERRORLOWER,
YERRORUPPER, GROUP, and INDEX.

SIZEMAX=dimension
starting with the second maintenance release of SAS 9.4, specifies a drawing size for
the largest marker when the marker size represents response values.
Note: Starting with the second maintenance release of SAS 9.4, the SIZEMAX=
option replaces the MARKERSIZEMAX= option. The syntax and functionality
are the same. The MARKERSIZEMAX= option is still honored, but the
SIZEMAX= option is preferred.

Default Three times the GraphDataDefault:MarkerSize setting (typically 7px)

or 21px.

Restriction The SIZEMAX= value must be greater than the SIZEMIN= or

MARKERSIZEMIN= value.

Interactions For this option to take effect, the SIZERESPONSE= or

MARKERSIZERESPONSE= option must also be used.

This option is ignored when the MARKERCHARACTER= option is

used. To manage the size of marker characters in a scatter plot, use the
TEXTPLOT statement.

Tip If you specify this size as a percent, then the specified value is
interpreted as a percent of the graph’s height. You can control the
height by using the DESIGNHEIGHT= option of the BEGINGRAPH
statement, or by using the HEIGHT= option of the ODS GRAPHICS
statement. For a standard 640px by 480px output size, a percentage
value of 4.5% sets a maximum size of about 21px, which is
approximately the same marker size that would result from this
option’s typical default setting.

See “dimension” on page 1340

SIZEMIN=dimension
starting with the second maintenance release of SAS 9.4, specifies a drawing size for
the smallest marker when the marker size represents response values.
SCATTERPLOT Statement 707

Note: Starting with the second maintenance release of SAS 9.4, the SIZEMIN=
option replaces the MARKERSIZEMIN= option. The syntax and functionality
are the same. The MARKERSIZEMIN= option is still honored, but the
SIZEMIN= option is preferred.

Default The GraphDataDefault:MarkerSize setting, which is typically 7px.

Restriction The SIZEMIN= value must be less than the SIZEMAX= or

MARKERSIZEMAX= value.

Interactions For this option to take effect, you must also specify the
SIZERESPONSE= or MARKERSIZERESPONSE= option.

This option is ignored when the MARKERCHARACTER= option is

used. To manage the size of marker characters in a scatter plot, use the
TEXTPLOT statement.

Tip If you specify this size as a percent, then the specified value is
interpreted as a percent of the graph’s height. You can control the
height by using the DESIGNHEIGHT= option of the BEGINGRAPH
statement, or by using the HEIGHT= option of the ODS GRAPHICS
statement. For a standard 640px by 480px output size, a percentage
value of 1.5% sets a minimum size of about 7px, which is
approximately the same marker size that would result from this
option’s typical default setting.

See “dimension” on page 1340

SIZERESPONSE=numeric-column | expression
starting with the second maintenance release of SAS 9.4, specifies a column that is
used to map the drawing size of the markers.
Note: Starting with the second maintenance release of SAS 9.4, the
SIZERESPONSE= option replaces the MARKERSIZERESPONSE= option. The
syntax and functionality are the same. The MARKERSIZERESPONSE= option
is still honored, but the SIZERESPONSE= option is preferred.
By default, the minimum and maximum values of this column establish a range over
which the marker sizes vary in linear proportion. The actual drawing size of the
smallest marker and the largest marker is set automatically.

Default The GraphDataDefault:MarkerSize setting, which is typically 7px.

Interactions This option overrides the SIZE= setting in the MARKERATTRS=

option.

This option is ignored when the MARKERCHARACTER= option is

used. To manage the size of marker characters in a scatter plot, use the
TEXTPLOT statement.

Tip You can adjust the smallest and largest marker size with the
SIZEMIN= and SIZEMAX= options, or with the
MARKERSIZEMIN= and MARKERSIZEMAX= options.

SUBPIXEL=AUTO | OFF
specifies whether subpixel rendering is used for image output when the scatter plot is
rendered.
708 Chapter 6 • Plot Statements

Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
AUTO
The system sets subpixel rendering OFF for this scatter plot, unless
SUBPIXEL=ON is specified in the BEGINGRAPH statement or in an ODS
GRAPHICS statement. In that case, subpixel rendering is ON for this scatter plot.
OFF
disables subpixel rendering for image output only for this scatter plot.

Default AUTO

Requirement Anti-aliasing must be enabled for this option to have any effect.

Notes This option affects subpixel rendering only for this plot. Subpixel
rendering for other plots in the graph is not affected.

For vector-graphics output, this option is ignored, and subpixel

rendering is always enabled.

When subpixel rendering is used for the graph but is turned OFF for
this scatter plot, some elements in the scatter plot such as the plot
markers might be offset a half pixel, which can make them appear
blurry in the image output.

Tips Anti-aliasing is enabled by default. If anti-aliasing has been disabled,

use the ANTIALIAS=ON option in the ODS GRAPHICS statement
to re-enable it.

Anti-aliasing is disabled automatically for this plot when the

resources required to anti-alias it exceed a preset threshold. When
anti-aliasing is disabled for this or any other plot in the graph,
subpixel rendering is disabled for the entire graph. A note is written
to the SAS log that provides information about how to use the
ANTIALIASMAX= option in an ODS GRAPHICS statement to re-
enable anti-aliasing.

See “Using Subpixel Rendering” in SAS Graph Template Language:

User's Guide

“ODS GRAPHICS Statement” in SAS ODS Graphics: Procedures

Guide for information about the ANTIALIAS=, ANTIALIASMAX=,
and SUBPIXEL= options.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the scatter
points. If this option is used, then it replaces all of the information that is displayed
by default. Roles for columns that do not contribute to the scatter plot can be
specified along with roles that do.
(role-list)
an ordered, space-separated list of unique SCATTERPLOT and user-defined
roles. SCATTERPLOT roles include X , Y , DATALABEL ,
MARKERCHARACTER , COLORRESPONSE or
MARKERCOLORGRADIENT , XERRORLOWER , XERRORUPPER ,
YERRORLOWER , YERRORUPPER , GROUP , and INDEX .
User-defined roles are defined with the ROLENAME= option.
SCATTERPLOT Statement 709

Requirement If you use the COLORRESPONSE= option, you must specify the
COLORRESPONSE role for the color values. Likewise, if you
use the MARKERCOLORGRADIENT= option, you must
specify the MARKERCOLORGRADIENT role for the color
values. Although they are functionally the same, you cannot mix
the COLORRESPONSE= and MARKERCOLORGRADIENT=
options, and their corresponding roles.

Example The following example displays data tips for the columns
assigned to the roles X, XERRORUPPER, and
XERRORLOWER, as well as the column Obs, which is not
assigned to any pre-defined SCATTERPLOT role. The Obs
column must first be assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X XERRORUPPER XERRORLOWER)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X , Y , DATALABEL ,
MARKERCHARACTER , COLORRESPONSE or
MARKERCOLORGRADIENT on page 702 , XERRORLOWER ,
XERRORUPPER , YERRORLOWER , YERRORUPPER , FREQ ,
and GROUP .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.
710 Chapter 6 • Plot Statements

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a point is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
marker that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable markers, you must include an ODS

GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding point is selected.

The URL value can be the same for any X and Y pairs. In that case,
the same action is taken when the points for those X and Y pairs are
selected.

USEDISCRETESIZE=TRUE | FALSE
specifies that the marker size should be based on fraction of the midpoint spacing
that is set by the DISCRETEMARKERSIZE= option.

Default FALSE

See DISCRETEMARKERSIZE=
SCATTERPLOT Statement 711

“boolean ” on page 1339 for other Boolean values that you can use.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interactions This option is ignored if the X= argument is not specified.

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

XERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the X error bars. The error bars are drawn
from the markers to the endpoints.

Default The lower segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the ERRORBARATTRS=

option.

If markers are displayed in the plot, then the markers overlay the error bars.
Large filled markers can obscure short error bars. To enable the error bars
to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the filled
markers.

XERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the X error bars. The error bars are drawn
from the markers to the endpoints.

Default The upper segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the ERRORBARATTRS=

option.

If markers are displayed in the plot, then the markers overlay the error bars.
Large filled markers can obscure short error bars. To enable the error bars
to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the filled
markers.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interactions This option is ignored if the Y= argument is not specified.

712 Chapter 6 • Plot Statements

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the Y error bars. The error bars are drawn
from the markers to the endpoints.

Default The lower segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the ERRORBARATTRS=

option.

If markers are displayed in the plot, then the markers overlay the error bars.
Large filled markers can obscure short error bars. To enable the error bars
to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the filled
markers.

YERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the Y error bars. The error bars are drawn
from the markers to the endpoints.

Default The upper segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the ERRORBARATTRS=

option.

If markers are displayed in the plot, then the markers overlay the error bars.
Large filled markers can obscure short error bars. To enable the error bars
to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the filled
markers.

Examples

Example 1: Grouped Scatter Plot

The following graph was generated by the “Example Program” on page 713:
Example 2: Discrete Scatter Plot 713

Example Program
proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Height and Weight by Sex";
layout overlay;
scatterplot x=height y=weight /
group=sex name="scatter" datalabel=name;
discretelegend "scatter";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.class template=scatterplot;

run;

Example 2: Discrete Scatter Plot

The following graph was generated by the “Example Program” on page 714:
714 Chapter 6 • Plot Statements

Example Program
/* Summarize SASHELP.CARS for mean MPG_HIGHWAY */
proc summary data=sashelp.cars nway;
class type;
var mpg_highway;
output out=mileage mean=mean;
run;

/* Sort by mileage */
proc sort data=mileage;
by mean;
run;

/* Define the graph template */

proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Average Highway MPG By Type";
layout overlay /
xaxisopts=(griddisplay=on gridattrs=(color=lightgray))
yaxisopts=(griddisplay=on gridattrs=(color=lightgray)
linearopts=(minorgrid=true minortickcount=9
minorgridattrs=(color=lightgray pattern=dot)));
scatterplot x=type y=mean /
group=type groupdisplay=cluster
markerattrs=graphDataDefault;
endlayout;
endgraph;
end;
run;
SCATTERPLOTMATRIX Statement 715

/* Render the graph */

proc sgrender data=mileage template=scatterplot;
run;

Details
This example creates a scatter plot of average highway mileage by vehicle type. The
SUMMARY procedure is used to compute the mean highway mileage for each vehicle
type. The SORT procedure is then used to sort the data by mileage in ascending order.
Normally, when you plot discrete values using the SCATTERPLOT statement, the
values at each end of the category axis are offset to accommodate the width of the largest
marker. When a small symbol is used, this can result in the minimum and maximum
symbols being placed very close to the edge of the axis. You can use the OFFSETMAX=
and OFFSETMIN= axis options to specify offsets for each end of the category axis.
However, you have to determine an appropriate offset value.
In this example, the GROUP= and GROUPDISPLAY= options are used to create offsets
automatically. The GROUP= option creates a group for each category value. The
GROUPDISPLAY=CLUSTER option displays the group values in a cluster and
automatically offsets the end values by one-half of the available midpoint spacing, which
is similar to the offsets that are used in bar charts. To restore the default colors and
symbols that are used for ungrouped data, the MARKERATTRS= option is used to set
the marker attributes to the graphDataDefault style element.

SCATTERPLOTMATRIX Statement
Creates a matrix of all pairwise scatter plots of the specified columns.
Restriction: The SCATTERPLOTMATRIX statement cannot appear within a LAYOUT OVERLAY,
LAYOUT OVERLAY3D, or LAYOUT OVERLAYEQUATED block. It is typically placed
in a LAYOUT GRIDDED block.

Syntax
SCATTERPLOTMATRIX numeric-column-list </option(s)>;

Summary of Optional Arguments

Appearance options
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
MARKERCOLORGRADIENT= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
starting with the second maintenance release of SAS 9.4, specifies the
column or range attribute map variable to use to determine the marker colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the plot markers.
DATATRANSPARENCY=number
specifies the degree of the transparency of the markers and data labels.
DIAGONAL=LABEL | (graph-list)
specifies whether the diagonal cells of the matrix are labeled with the labels
(names) of the required arguments, or with a plot.
INDEX=positive-integer-column | expression
716 Chapter 6 • Plot Statements

specifies indices for mapping marker attributes (color and symbol) to one of
the GraphData1–GraphDataN style elements.
LABELSTRIP=TRUE | FALSE
specifies whether leading and trailing blanks are stripped from marker
characters or data labels that have a fixed position before they are displayed
in the plot.
MARKERATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the data markers.
MARKERCHARACTER=column | expression
specifies a column that defines strings that are to be used instead of marker
symbols.
MARKERCHARACTERATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the marker character specified on the
MARKERCHARACTER= option.
MARKERCHARACTERPOSITION=CENTER | TOP | BOTTOM | LEFT | RIGHT |
TOPLEFT | TOPRIGHT | BOTTOMLEFT | BOTTOMRIGHT
specifies the justification of the marker characters.
MARKERCOLORGRADIENT=numeric-column | range-attr-var | expression
in the first maintenance release of SAS 9.4 and earlier releases, specifies the
column or range attribute map variable that is used to determine the marker
colors.
MATRIXTYPE=FULL | UPPERTRIANGLE | LOWERTRIANGLE
specifies whether to display the full matrix, or just the upper or lower triangle
of the matrix.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.
START=TOPLEFT | BOTTOMLEFT
specifies whether to start populating the matrix from the top left or bottom
left corner.
SUBPIXEL=AUTO | OFF
specifies whether subpixel rendering is used for image output when the
scatter plots are rendered.
WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
WALLDISPLAY=STANDARD | ALL | NONE | (display-options)
specifies whether the plot’s wall and wall outline are displayed.

Confidence options
ELLIPSE=(<ellipse-suboptions>)
specifies that a confidence ellipse be included in each cell containing a scatter
plot.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
scatter points.
TIPFORMAT=(role-format-list)
SCATTERPLOTMATRIX Statement 717

specifies display formats for tip columns.

TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Inset options
INSET=(info-options)
specifies what information is displayed in an inset.
INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information.

Label options
DATALABEL=column
specifies a column for marker labels.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |
CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the markers.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label blocks.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of scatter markers, error bars, and data labels for each
unique group value of the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
CORROPTS=(correlation-options)
specifies options for computing measures of association between pairs of
columns.
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation
that is read.
ROWVARS=(numeric-column-list)
specifies a secondary list of numeric columns to be paired with the required
column list that is specified by the SCATTERPLOTMATRIX statement.
718 Chapter 6 • Plot Statements

Required Argument
numeric-column-list
specifies a list of numeric columns to plot.

Requirements There must be at least two columns to produce a useful matrix.

All of the columns must be numeric.

Note The default width is 640px, and the default height is 480px. The
graph size is not automatically adjusted to accommodate a large
number of columns.

Tips To change the graph size for the current template, use the
DESIGNHEIGHT= and DESIGNWIDTH= options in the
BEGINGRAPH statement.

To change the graph size for all templates in the current SAS
session, use the HEIGHT= and WIDTH= options in the ODS
GRAPHICS statement. Size settings in the ODS GRAPHICS
statement override size settings in the BEGINGRAPH statement.

You can also limit the number of columns in the matrix (perhaps to
seven in each dimension, for example) so that the resulting graphs
are not too small to be useful.

Optional Arguments
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
MARKERCOLORGRADIENT= option.
style-element
specifies the name of a style element. The style element should contain these
style attributes:

STARTCOLOR specifies a color for the smallest data value of the

COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
NEUTRALCOLOR specifies a color for the midpoint of the range of the
COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
ENDCOLOR specifies a color for the highest data value of the
COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

SCATTERPLOTMATRIX Statement 719

Default The ThreeColorAltRamp style element.

Interactions For this option to take effect, the COLORRESPONSE= or

MARKERCOLORGRADIENT= option must also be specified.

The REVERSECOLORMODEL= option can be used to reverse the

start and end colors of the ramp assigned to the color model.

COLORRESPONSE=numeric-column | range-attr-var | expression

starting with the second maintenance release of SAS 9.4, specifies the column or
range attribute map variable to use to determine the marker colors.
Note: Starting with the second maintenance release of SAS 9.4, the
COLORRESPONSE= option replaces the MARKERCOLORGRADIENT=
option. The syntax and functionality are the same. The
MARKERCOLORGRADIENT= option is still honored, but the
COLORRESPONSE= option is preferred.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

Tip For each range in the attribute map, the RANGEALTCOLOR= or

RANGEALTCOLORMODEL= option in the RANGE statement
determines the marker colors.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.
You can use this option to add a second response variable to an analysis. For
example, in an analysis of weight by height, you can specify an age column by using
the COLORRESPONSE= or MARKERCOLORGRADIENT= option so that the
change in the gradient color of the markers reflects the change in age.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column has
720 Chapter 6 • Plot Statements

multiple values for a single GROUP value, unexpected results might

occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

Suboption COLOR= in the DATALABELATTRS= option overrides

this option for the data label color attribute.

Suboption COLOR= in the MARKERATTRS= option or in the

MARKERCHARACTERATTRS= option overrides this option for
the marker colors.

Note You can use MARKERCOLORGRADIENT= as an alternative to

COLORRESPONSE=. However, if you use
MARKERCOLORGRADIENT=, be aware that the TIP=,
TIPFORMAT=, and TIPLABEL= options recognize
MARKERCOLORGRADIENT as the color role, not
COLORRESPONSE.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

If the MARKERCHARACTER= option is also specified, then the

gradients that would be applied to the markers are applied to the text
strings.

CORROPTS=(correlation-options)
specifies options for computing measures of association between pairs of columns.
The following correlation-options are available:
EXCLNPWGT=TRUE | FALSE
specifies whether observations with non-positive weight values are excluded
(TRUE) from the analysis.

Default FALSE (observations with negative weights are treated like those with
zero weights and counted in the total number of observations).

See “boolean ” on page 1339 for other Boolean values that you can use.

NOMISS=TRUE | FALSE
specifies whether observations with missing values are excluded (TRUE) from
the analysis.

Default FALSE (correlation statistics are computed using all of the nonmissing
pairs of columns).

Note Using NOMISS=TRUE is computationally more efficient.

See “boolean ” on page 1339 for other Boolean values that you can use.

WEIGHT=numeric-column
specifies a weighting column to use in the calculation of Pearson weighted
product-moment correlation.
The observations with missing weights are excluded from the analysis. If you use
this WEIGHT correlation option, then consider which value of the
VARDEF=correlation option is appropriate.
SCATTERPLOTMATRIX Statement 721

Default For observations with non-positive weights, the weights are set to zero
and the observations are included in the analysis.

Tip You can include EXCLNPWGT among the correlation options to

exclude observations with negative or zero weights from the analysis.

VARDEF=DF | N | WDF | WEIGHT

specifies the variance divisor in the calculation of variances and covariances.

DF Degrees of Freedom (N – 1)
N number of observations
WDF sum of weights minus 1 (WEIGHT – 1)
WEIGHT sum of weights

Default DF

Interaction This option has effect only when the INSET= option is also used.

See the CORR procedure information in Base SAS Procedures Guide:

Statistical Procedures for the statistical and computational details of
these options.

DATALABEL=column
specifies a column for marker labels. The label positions are adjusted to prevent the
labels from overlapping.

Interactions If a numeric column is specified and the column has no format, then a
BEST6 format is applied.

This option is ignored if the MARKERCHARACTER= option is used.

Note The position of the labels are adjusted to prevent the labels from
overlapping.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the DATALABEL= option must also be
specified.

This option is ignored if the MARKERCHARACTER= option is

specified.

Note When the DATALABELPOSITION=AUTO option is in effect, in

some cases, the data label font size might be reduced in order to avoid
overlapping labels and markers.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

722 Chapter 6 • Plot Statements

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |

CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the markers.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
data label. In that case, all of the specified split characters together are treated as a
single split character.
When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is
split unconditionally at each occurrence of any of the specified split characters. If the
data label does not contain any of the specified characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

SCATTERPLOTMATRIX Statement 723

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
TRUE
drops the split characters from the data label.
FALSE
includes the split characters in the data label. When DATALABELSPLIT=TRUE
and DATALABELSPLITCHARDROP=FALSE, each split character remains as
the last character in the current line. The characters that follow the split character,
up to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a data label with the following
specifications:
• the data label text for this label is Product*Group*A
• DATALABELSPLIT=TRUE
• DATALABELSPLITCHARDROP=TRUE | FALSE
• DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from the

label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains as the
last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.
AUTO
justifies the labels based on the DATALABELPOSITION= option, as shown in
the following table.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

724 Chapter 6 • Plot Statements

DATALABELPOSITION= Value Justification

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which DATALABELPOSITION=TOP.
Note: The gray vertical line at the bottom of each label represents the horizontal
center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of text.

The text box is anchored the same way that the unsplit text is anchored. For example,
if DATALABELPOSITION=TOP, then the bottom center of the text box is
positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if DATALABELSPLIT=FALSE, or if

DATALABELSPLIT=TRUE and DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the plot markers. The following figure shows large
CIRCLEFILLED markers with each of the skins applied.
SCATTERPLOTMATRIX Statement 725

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the GraphSkins:DataSkin
style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

When a data skin is applied, all marker outlines are set by the skin,
and the outline settings from the ODS style or from the
MARKERATTRS= option are ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the markers and data labels.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DIAGONAL=LABEL | (graph-list)
specifies whether the diagonal cells of the matrix are labeled with the labels (names)
of the required arguments, or with a plot. The plot for each diagonal cell consists of
an overlay combination of a histogram, normal, or kernel curves.
The graph-list is a space-separated list of one or more of the following:

HISTOGRAM specifies a histogram

NORMAL specifies a normal density curve
KERNEL specifies a kernel density estimate.

Default LABEL. Column labels (or names) are displayed in the diagonal
cells.

Requirement When specifying multiple graphs in the graph-list, you must separate
the values with a space. For example, the following specification
requests both a histogram and a normal density curve in each
diagonal cell:

DIAGONAL=(HISTOGRAM NORMAL)

Interactions The computation for HISTOGRAM, NORMAL, and KERNEL is

always computed on all the data for the current column (including the
FREQ= column, if used). The GROUP= option is not considered in
any of these computations.

This option is ignored if the ROWVARS= option is used.

Note When this option is specified, the labels are drawn around the outside
of the matrix, and the matrix axes are dropped.
726 Chapter 6 • Plot Statements

ELLIPSE=(<ellipse-suboptions>)
specifies that a confidence ellipse be included in each cell containing a scatter plot.
The ellipse is always drawn behind the scatter points. The ellipse-suboptions include
the following:
ALPHA=positive-number
specifies the confidence level to compute for each ellipse.

Default 0.05

Range 0 < number < 1

Tip ALPHA=0.05 represents a 95% confidence level.

CLIP=TRUE | FALSE
specifies whether the X and Y values for the ellipse are considered when
determining the data ranges for the axes.
TRUE
the X and Y values for the ellipses are ignored when the axis ranges are
determined. Clipping occurs if the X and Y values for an ellipse exceed the
axis range.
FALSE
the X and Y values for the ellipses contribute to the data range for each axis.
If necessary, each axis is extended in order to display the entire ellipse.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

TYPE=MEAN | PREDICTED
specifies the type of ellipse.

MEAN specifies a confidence ellipse of the mean

PREDICTED specifies a prediction ellipse of the data

Default MEAN

See “ELLIPSE Statement” on page 422 for statistical details about how the
ellipse is calculated.

Default TYPE=MEAN ALPHA=0.05 You can set defaults by specifying the

option without arguments: ELLIPSE=( ).

Interactions The ellipse might be clipped by the data range for the scatter points.

The ellipse is always computed on all the data for the current pair of X
and Y columns (including the FREQ= column, if used). The GROUP=
option is not considered when computed the ellipse.

Tip The display properties of each ellipse are controlled by the style
elements. The GraphDataDefault element controls the outline and fill
properties, and the GraphEllipse element controls whether the outline,
fill, or both are shown.
SCATTERPLOTMATRIX Statement 727

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that is
read.

Default All observations have a frequency count of 1.

Restriction If the value of the numeric-column is missing or is less than 1, then the
observation is not used in the analysis. If the value is not an integer,
then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation.

GROUP=column | discrete-attr-var | expression

creates a distinct set of scatter markers, error bars, and data labels for each unique
group value of the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the graph by a

different combination of color and marker symbol. Markers vary
according to the ContrastColor and MarkerSymbol attributes of the
GraphData1–GraphDataN and GraphMissing style elements.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of markers and
colors.

The marker size is set by the MARKERATTRS= option.

The MARKERCHARACTER= , COLORRESPONSE= , and

MARKERCOLORGRADIENT= options override the group settings
for the marker symbol and marker color.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Tip The representations that are used to identify the groups can be
overridden. For example, each distinct group value is represented by a
different marker symbol, but the
MARKERATTRS=(SYMBOL=marker) option could be used to
assign the same symbol to all of the plot’s marker symbols, letting
marker color indicate group values. Likewise,
MARKERATTRS=(COLOR=color) could be used to assign the same
color to all markers, letting marker symbol indicate group values.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.
728 Chapter 6 • Plot Statements

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping marker attributes (color and symbol) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interactions For this option to take effect, the GROUP= option must also be
specified.

The MARKERCHARACTER=, COLORRESPONSE=, and

MARKERCOLORGRADIENT= options override the group settings
for the marker symbol and marker color.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

INSET=(info-options)
specifies what information is displayed in an inset. Insets appear in all cells of the
matrix except the diagonal and are displayed as a small table of name-value pairs.
The following info-options are available:
NOBS
displays the total number of observations where both the X and Y columns have
nonmissing values. If the FREQ= option is used, then this number is adjusted
accordingly. The value of NOBS can be further adjusted by the use of the
SCATTERPLOTMATRIX Statement 729

NOMISS=, WEIGHT=, and EXCLNPWGT= suboptions of the CORROPTS=

option.
PEARSON
displays the Pearson product-moment correlation. The computation of the
correlation is affected by the FREQ= and CORROPTS= options. The
computation is not done on a per group value when GROUP= is used.
PEARSONPVAL
displays the probability value for the Pearson product-moment correlation.

Tip The location and appearance of the inset is controlled by the

INSETOPTS= option.

See PROC CORR in the documentation for Base SAS for statistical and
computational details of these options.

Example Here is an example of a typical inset:

N 150
r 0.96287
p(r) <.0001
In this example, NOBS is represented by N, PEARSON is represented by
r, and PEARSONPVAL is represented by p(r).

INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information. The appearance
options can be any one or more of the settings that follow. The options must be
enclosed in parentheses, and each option is specified as a name=value pair.
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the inset is automatically aligned within the layout.
NONE
does not automatically align the inset. The inset’s position is therefore set by
the HALIGN= and VALIGN=appearance-options.
AUTO
attempts to center the inset in the area that is farthest from any surrounding
markers. Data cells might have different inset placements.
(location-list)
restricts the inset’s possible locations to those locations in the specified
location-list, and uses the location-list position that least collides with the
data cell’s other graphics features. The location-list is space-separated and
can contain any of these locations: TOPLEFT TOP TOPRIGHT LEFT
CENTER RIGHT BOTTOMLEFT BOTTOM BOTTOMRIGHT. Example:
AUTOALIGN=(TOPRIGHT TOPLEFT)

Default NONE

Interaction When AUTOALIGN=AUTO or (location-list), the enclosing layout

statement’s HALIGN= and VALIGN=appearance-options are
ignored.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the inset background
730 Chapter 6 • Plot Statements

style-reference
specifies a style reference in the form style-element : style-attribute. Only the
style-attribute named COLOR or CONTRASTCOLOR is used.

Default GraphWalls:Color style reference

BORDER=TRUE | FALSE
specifies whether a border is displayed around the inset.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

HALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment of the inset.

Default LEFT

Interaction This option is ignored when AUTOALIGN= is not NONE and the
parent layout is an overlay-type layout.

OPAQUE=TRUE | FALSE
specifies whether the inset background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the entire inset.

Default The GraphDataText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TITLE="string"
specifies a title for the inset. The title is added at the top of the inset and spans
the full inset width.

Note Space is not reserved for the title when this option is not set

Tip Text properties for the title string can be set with TITLEATTRS=.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the inset’s title string.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

SCATTERPLOTMATRIX Statement 731

VALIGN=TOP | CENTER |BOTTOM

specifies the vertical alignment of the inset.

Default TOP

Interaction This option is ignored when AUTOALIGN= is not NONE and the
parent layout is an overlay-type layout.

LABELSTRIP=TRUE | FALSE
specifies whether leading and trailing blanks are stripped from marker characters or
data labels that have a fixed position before they are displayed in the plot. The
MARKERCHARACTER= option specifies the column that provides the marker
strings that are to be used in place of marker symbols.

Default FALSE

Interactions This option effects marker strings only when the

MARKERCHARACTER= option is specified.

This option effects data labels only when DATALABEL= is specified

and DATALABELPOSITION= is not AUTO.

Tip Stripping the blanks from the numeric value strings helps center each
string relative to its data point. Stripping is useful when you want to
overlay the data values near or inside the markers for a plot.

See “boolean ” on page 1339 for other Boolean values that you can use.

MARKERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes of

the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions If the COLORRESPONSE= or MARKERCOLORGRADIENT=

option is specified, then this option’s COLOR= suboption is ignored.

If the MARKERCHARACTER= option is specified, then this option’s

SYMBOL= and WEIGHT= suboptions are ignored.

This option’s COLOR= suboption overrides the default behavior for

grouped data. When the COLOR= suboption is specified in that case,
all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior for

grouped data. When the SYMBOL= suboption is specified in that
case, all markers have the same symbol, and the symbol color alone
distinguishes the markers.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.
732 Chapter 6 • Plot Statements

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Marker Options” on page 1350 for available marker-options.

MARKERCHARACTER=column | expression
specifies a column that defines strings that are to be used instead of marker symbols.

Each string is centered horizontally and vertically at the data point. The data point
positions are not adjusted to prevent text overlap.

Interactions This option overrides the DATALABEL= option.

If the GROUP= option is also used, then color is displayed for a

DISCRETE legend, but the character is not displayed in the legend.
For text strings, the same colors are applied to the text strings as
would have been applied to markers.

Note If a numeric column is used, then its values are converted to strings
using the format associated with the column or BEST6 if no format is
defined.

Tips Lengthy strings might be clipped by the plot border. To reduce

clipping, you can use the OFFSETMIN= and OFFSETMAX=
suboptions of the XAXISOPTS= and YAXISOPTS= options to
increase the axis offsets.

You can use the MARKERCHARACTERPOSITION= option to

change the justification of the marker character.

You can use the LABELSTRIP= option to strip the leading and
trailing blanks from numeric value strings in order to center each
string on its data point.

MARKERCHARACTERATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the marker character specified on the
MARKERCHARACTER= option.
SCATTERPLOTMATRIX Statement 733

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the MARKERCHARACTER= option

must also be used.

When the GROUP= option is also specified, each distinct group value
might be represented by a different color (depending on the ODS style
setting or the setting on the INDEX= option). The marker character
that is associated with the group is assigned the group color.

This option’s COLOR= suboption can be used to specify a single color

for all marker characters in a grouped plot, without affecting items that
have a group color, such as error bars and marker symbols.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

MARKERCHARACTERPOSITION=CENTER | TOP | BOTTOM | LEFT |

RIGHT | TOPLEFT | TOPRIGHT | BOTTOMLEFT | BOTTOMRIGHT
specifies the justification of the marker characters. The following figure shows the
effect of each of the values on the position of marker character M.

Note: The red dots represent the center in each case for reference.

Default CENTER

Interaction This option is ignored if the MARKERCHARACTER= option is not

specified.

Tip You can use the LABELSTRIP= option to strip the leading and trailing
blanks from value strings in order to properly justify each string on its
data point.

MARKERCOLORGRADIENT=numeric-column | range-attr-var | expression

in the first maintenance release of SAS 9.4 and earlier releases, specifies the column
or range attribute map variable that is used to determine the marker colors.
Note: Starting with the second maintenance release of SAS 9.4, the
MARKERCOLORGRADIENT= option is deprecated and replaced with the
COLORRESPONSE= option. The syntax and functionality are the same. The
734 Chapter 6 • Plot Statements

MARKERCOLORGRADIENT= option is still honored, but the

COLORRESPONSE= option is preferred.

Note Starting with the second maintenance release of SAS 9.4, if you use the
MARKERCOLORGRADIENT= option, then be aware that the TIP=,
TIPFORMAT=, and TIPLABEL= options recognize the
MARKERCOLORGRADIENT role and not the COLORRESPONSE role.

See COLORRESPONSE= on page 719

MATRIXTYPE=FULL | UPPERTRIANGLE | LOWERTRIANGLE

specifies whether to display the full matrix, or just the upper or lower triangle of the
matrix. By default, the full matrix is displayed. The cells in the grid are filled
beginning with the cell that is specified by the START= option. When you display
only the upper or lower triangle, you can use the START= option to control the
orientation of the triangle. The following figure shows the effect of the
MATRIXTYPE= option when START=TOPLEFT and
DIAGONAL=(HISTOGRAM).

The next figure shows the effect of the START=BOTTOMLEFT option on the
previous graph.

Default FULL

Interactions This option is ignored if the ROWVARS= option is used.

The START= option specifies the corner where the matrix fill begins.
SCATTERPLOTMATRIX Statement 735

The DIAGONAL= option specifies the content of the diagonal cells.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either the
ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, DATALABEL, MARKERCHARACTER,
COLORRESPONSE or MARKERCOLORGRADIENT, GROUP,
and INDEX.

ROWVARS=(numeric-column-list)
specifies a secondary list of numeric columns to be paired with the required column
list that is specified by the SCATTERPLOTMATRIX statement. The labels for the
columns appear vertically on the left side of the matrix.

Requirement All of the columns must be numeric.

Interaction When this option is specified, the DIAGONAL= and

MATRIXTYPE= options are ignored.

START=TOPLEFT | BOTTOMLEFT
specifies whether to start populating the matrix from the top left or bottom left
corner.

Default TOPLEFT
736 Chapter 6 • Plot Statements

SUBPIXEL=AUTO | OFF
specifies whether subpixel rendering is used for image output when the scatter plots
are rendered.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
AUTO
The system sets subpixel rendering OFF for this scatter-plot matrix, unless
SUBPIXEL=ON is specified in the BEGINGRAPH statement or in an ODS
GRAPHICS statement. In that case, subpixel rendering is ON for this scatter-plot
matrix.
OFF
disables subpixel rendering for image output only for the scatter plots in this
scatter-plot matrix.

Default AUTO

Restriction This option applies only to the scatter plots in the matrix. It does not
affect the confidence ellipses.

Requirement Anti-aliasing must be enabled for this option to have any effect.

Notes This option affects subpixel rendering for the scatter plots only in this
scatter-plot matrix. Subpixel rendering for other plots in the graph is
not affected.

For vector-graphics output, this option is ignored, and subpixel

rendering is always enabled.

When subpixel rendering is used for the graph but is turned OFF for
this scatter-plot matrix, some elements in the scatter plots such as the
plot markers might be offset a half pixel, which can make them
appear blurry in the image output.

Tips Anti-aliasing is enabled by default. If anti-aliasing has been disabled,

use the ANTIALIAS=ON option in the ODS GRAPHICS statement
to re-enable it.

Anti-aliasing is disabled automatically for this plot when the

resources required to anti-alias it exceed a preset threshold. When
anti-aliasing is disabled for this or any other plot in the graph,
subpixel rendering is disabled for the entire graph. A note is written
to the SAS log that provides information about how to use the
ANTIALIASMAX= option in an ODS GRAPHICS statement to re-
enable anti-aliasing.

To disable subpixel rendering for the scatter plots and the ellipses,
specify SUBPIXEL=OFF in the template’s BEGINGRAPH statement
or in an ODS GRAPHICS statement.

See “Using Subpixel Rendering” in SAS Graph Template Language:

User's Guide

“ODS GRAPHICS Statement” in SAS ODS Graphics: Procedures

Guide for information about the ANTIALIAS=, ANTIALIASMAX=,
and SUBPIXEL= options.
SCATTERPLOTMATRIX Statement 737

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the scatter
points. If this option is used, then it replaces all of the information that is displayed
by default. Roles for columns that do not contribute to the scatter plot can be
specified along with roles that do.
(role-list)
an ordered, space-separated list of unique SCATTERPLOTMATRIX and user-
defined roles. SCATTERPLOT roles include: X, Y, GROUP, DATALABEL,
MARKERCHARACTER, and COLORRESPONSE or
MARKERCOLORGRADIENT.
User-defined roles are defined with the ROLENAME= option.

Requirement If you use the COLORRESPONSE= option, you must specify the
COLORRESPONSE role for the color values. Likewise, if you
use the MARKERCOLORGRADIENT= option, you must
specify the MARKERCOLORGRADIENT role for the color
values. Although they are functionally the same, you cannot mix
the COLORRESPONSE= and MARKERCOLORGRADIENT=
options, and their corresponding roles.

Example The following example displays data tips for the columns
assigned to the roles TIP1, TIP2, TIP3, and TIP4:

ROLENAME=(TIP1=ID TIP2=AGE TIP3=HEIGHT TIP4=WEIGHT)

TIP=(TIP1 TIP2 TIP3 TIP4)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: current X, current Y, DATALABEL,
MARKERCHARACTER, COLORGROUP or
MARKERCOLORGRADIENT, and GROUP.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
738 Chapter 6 • Plot Statements

TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style- attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.
STANDARD
displays a filled wall. The setting of the FRAMEBORDER=ON | OFF attribute
of the GraphWalls style element determines whether the wall outline is displayed.
ALL
displays a filled, outlined wall.
NONE
displays no wall, no wall outline.
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays the wall outline.

FILL displays a filled wall area.
Example: SCATTERPLOTMATRIX Statement 739

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall outline are set by the GraphAxisLine
style element.

Details
By default, the SCATTERPLOTMATRIX statement produces a symmetric scatter plot
matrix. For n columns, it produces an n columns by n rows matrix of scatter plots. By
default, the columns of the matrix are in the same left-to-right order as the order of the
numeric-column-list. The rows of the matrix are in the same bottom-to-top order as the
numeric-column-list. You can reverse the direction of the diagonal by setting
START=TOPLEFT.
To produce a rectangular matrix of scatter plots, use the ROWVARS= option. Specifying
n columns in the SCATTERPLOTMATRIX statement and m columns on the
ROWVARS= option produces an n-columns by m-rows matrix of scatter plots. For
example, the following statement specifies 2 columns on SCATTERPLOTMATRIX and
3 columns on the ROWVARS= option to produce the 2-columns by 3-rows matrix:

SCATTERPLOTMATRIX Height Weight

/ ROWVARS=(Age Height Weight);

The SCATTERPLOTMATRIX statement cannot appear within an overlay-type layout. It

generates its own matrix of plots and is typically placed in a LAYOUT GRIDDED
block.
If there are missing values in a column or a row, then all of the points that can be plotted
are plotted in each scatter plot.

Example: SCATTERPLOTMATRIX Statement

The following graph was generated by the “Example Program” on page 740:
740 Chapter 6 • Plot Statements

Example Program
proc template;
define statgraph scatterplotmatrix;
begingraph;
entrytitle "Scatter Plot Matrix";
layout gridded;
scatterplotmatrix
sepallength sepalwidth petallength petalwidth /
group=species name="matrix";
discretelegend "matrix";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.iris template=scatterplotmatrix;

run;

SERIESPLOT Statement
Displays a series of line segments that connect observations of input data.
Tip: Starting with the third maintenance release of SAS 9.4, subpixel rendering is enabled
by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page
33.For information about the ODS GRAPHICS statement SUBPIXEL= option, see
“ODS GRAPHICS Statement” in SAS ODS Graphics: Procedures Guide.
SERIESPLOT Statement 741

Syntax
SERIESPLOT X=column | expression
Y=column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
ARROWHEADPOSITION= NONE | START | END | BOTH | column
specifies a position for arrowheads.
ARROWHEADSCALE=positive-number | numeric-column | expression
specifies an arrowhead scale factor based on the thickness of the arrow line.
ARROWHEADSHAPE= OPEN | FILLED | BARBED | column
specifies a shape for arrowheads.
BREAK=TRUE | FALSE
breaks the plot line at missing values of the X or Y column.
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group
clusters as a fraction of the minimum interval between adjacent data values.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the line and
marker colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the series lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the line, markers, line label, and
data labels, when displayed.
DISPLAY=STANDARD | ALL | display-options
specifies additional feature to display with the series line.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line
pattern) to one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the series line.
LINECOLORGROUP=column | discrete-attr-var | expression
specifies a column that determines the line colors for a grouped plot
independently of the GROUP= column.
LINEPATTERNGROUP=column | discrete-attr-var | expression
specifies a column that determines the line patterns for a grouped plot
independently of the GROUP= column.
LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line thickness.
LINETHICKNESSMIN=dimension
742 Chapter 6 • Plot Statements

specifies the minimum line thickness when a response variable is used to

determine the line thickness.
LINETHICKNESSRESPONSE=numeric-column | range-attr-var | expression
specifies a response column or range attribute variable that is used to map a
line thickness to each group value.
MARKERATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the data markers.
MARKERCOLORGROUP=column | discrete-attr-var | expression
specifies a column that determines the marker colors for a grouped plot
independently of the GROUP= column.
MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.
MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.
MARKERSYMBOLGROUP=column | discrete-attr-var | expression
specifies a column that determines the marker symbols for a grouped plot
independently of the GROUP= column.
SMOOTHCONNECT=TRUE | FALSE
specifies that a smoothed line passes through all vertices.
SPLINEPOINTS=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the
INTERVAL= axis option.
SPLINETYPE=NONE | QUADRATICBEZIER
specifies the type of spline interpolation that is used to draw the series line.

Axes options
CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Connect options
CONNECTORDER=XVALUES | XAXIS
specifies how to connect the data points to form the series line.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
series line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
SERIESPLOT Statement 743

TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string" | column | expression
specifies a label for the series line.
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the series-line labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the series-line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the series-line label relative to the series line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the series-line label at the specified split characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the series-line label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the series-line label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the series-line label
block.
DATALABEL=column | expression
specifies a column that supplies values for the data point labels.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |
CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the vertices of the series
line and the markers, when displayed.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all series lines and markers from discrete X
values, or discrete Y values, or both.
GROUP=column | discrete-attr-var | expression
creates a separate series plot for each unique group value in the specified
column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.
744 Chapter 6 • Plot Statements

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a point or a segment of the curve is
selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies the column for the X values.
Y=column | expression
specifies the column for the Y values.

Optional Arguments
ARROWHEADPOSITION= NONE | START | END | BOTH | column
specifies a position for arrowheads.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

NONE no arrowheads are displayed

START an arrowhead is displayed at the starting point of each line
END an arrowhead is displayed at the ending point of each line.
BOTH an arrowhead is displayed at each end of each line.
column specifies a column that provides an arrowhead position for each group
value.

Default NONE

Restriction When you specify a columnand the GROUP= option is in effect, the
arrowhead position values are assumed to be constant for each group
value. If the column has multiple values for a single group value, then
only one of the arrowhead position values is used for that group.

See “Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

ARROWHEADSCALE=positive-number | numeric-column | expression

specifies an arrowhead scale factor based on the thickness of the arrow line.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
SERIESPLOT Statement 745

Default 1.0

Restriction The arrowhead scale values are assumed to be constant for each line. If
a column or expression provides multiple values for a line, only one of
the values is used.

Interaction This option is ignored when ARROWHEADPOSITION=NONE is in

effect.

Tip Use a factor greater than 1.0 to make a larger arrowhead.

See “Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

ARROWHEADSHAPE= OPEN | FILLED | BARBED | column

specifies a shape for arrowheads.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
The following figure shows each of the arrowhead shapes.

Default OPEN

Restriction When you specify a column and the GROUP= option is in effect, the
arrowhead shape values are assumed to be constant for each group
value. If the column has multiple values for a single group value, only
one of the arrowhead shape values is used for that group.

Interaction This option is ignored when ARROWHEADPOSITION=NONE is in

effect.

See “Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

BREAK=TRUE | FALSE
breaks the plot line at missing values of the X or Y column.

Default FALSE

Note When this option is set to FALSE, missing values are skipped and a
continuous line is drawn.

See “boolean ” on page 1339 for other Boolean values that you can use.

CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when GROUPDISPLAY=CLUSTER.

AUTO uses the discrete axis for clustering groups when only one axis is
discrete. Uses the X axis for clustering if both axes are discrete or
interval.
X|Y uses the X or Y axis for clustering groups.
746 Chapter 6 • Plot Statements

Default AUTO

Interaction The GROUPDISPLAY= option must be set to CLUSTER for this

option to have any effect.

CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters as a
fraction of the minimum interval between adjacent data values.

Default 0.85

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.

Interactions For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

When markers are displayed for interval data and

GROUPDISPLAY=CLUSTER and CLUSTERWIDTH= are in effect,
the size of the markers in each cluster might be reduced to no less than
5 pixels in order to display the cluster within the smallest effective
midpoint space. If you need larger markers in that case, use the
MARKERATTRS= option to specify a larger marker size.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.
SERIESPLOT Statement 747

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the line and marker
colors.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Requirement The COLORRESPONSE values should remain constant for each

group value in a grouped plot and for the entire plot in an ungrouped
plot. If the COLORRESPONSE column has multiple values for a
single GROUP value or for a non-grouped plot, unexpected results
might occur.

Interactions When this option is specified without the GROUP= option, only a
single line is generated for the plot, and the line color is derived from
the COLORRESPONSE= value.

When the GROUP= option, LINECOLORGROUP= option, and

MARKERCOLORGROUP= options are used with the
COLORRESPONSE= option, the COLORRESPONSE= option
controls the color attributes.

Suboption COLOR= in the DATALABELATTRS= option overrides

this option for the data label color attribute.

This option overrides suboption COLOR= in the LINEATTRS=

option and varies the line color according to the color gradient or the
attribute map.

Suboption COLOR= in the MARKERATTRS= option or in the

FILLEDOUTLINEDMARKERS= option overrides this option for
the marker colors.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

For a numeric column or expression, the ThreeColorAltRamp style

element defines the line color gradient.

CONNECTORDER=XVALUES | XAXIS
specifies how to connect the data points to form the series line.
XVALUES
connects data points in the order read from the X column.
748 Chapter 6 • Plot Statements

XAXIS
connects data points as they occur min-to-max along the X axis.

Default XVALUES

Tip For certain types of series lines (for example, time series) when the input
data might not be sorted by the X column, set this option to XAXIS to
assure the expected connect order.

CURVELABEL="string" | column | expression

specifies a label for the series line.

Default No series-line label is displayed

Restrictions When the GROUP= option is specified, "string" and expression are
not valid. Use column in that case.

When the GROUP= option is not specified, column is not valid. Use
"string" or expression in that case.

The line label for missing values is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

See GROUP= on page 758

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the series-line labels. When the GROUP=
option is used, each distinct group value might be represented by a different color.
The series-line label that is associated with the group is assigned the group color.
This option can be used to specify a single color for all series-line labels in a plot,
without affecting items that have the group color, such as lines and marker symbols.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, text color is derived from the

GraphData1:ContrastColor–GraphDataN:ContrastColor style
references. The font is derived from the GraphValueText style
element.

Interactions For this option to take effect, the CURVELABEL= option must also
be used.

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the series-line label relative to the plot area.
INSIDE
locates the labels inside the plot area
SERIESPLOT Statement 749

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the SERIESPLOT is used in multicell

layouts such as LATTICE, DATAPANEL, or DATALATTICE, where
axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the series-line
labels appear. For more information, see “Location and Position of
Curve Labels” on page 185.

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the series-line label relative to the series line.
AUTO
automatically locates the series-line label near the end series line along unused
axes whenever possible (typically Y2 or X2) in order to avoid collision with tick
values.

Restriction This option is used only when CURVELABELLOCATION=

OUTSIDE.

MAX
forces the series-line label to appear near maximum series values (typically, to
the right)
MIN
forces the series-line label to appear near minimum series values (typically, to the
left)
START
forces the series-line label to appear near the beginning of the series line.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the series line has a spiral
shape.

END
forces the series-line label to appear near the end of the series line.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the series line has a spiral
shape.

Defaults AUTO when CURVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

750 Chapter 6 • Plot Statements

Restriction The AUTO setting is ignored if CURVELABELLOCATION=INSIDE

is specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the series-
line label appears. For more information, see “Location and Position
of Curve Labels” on page 185.

Note When you specify TICKVALUELIST=, VIEWMAX=, or

VIEWMIN= in an axis statement, the data points that are used to
determine the position of the series-line label might fall outside of the
graph area. In that case, the series-line label might not be displayed or
might be positioned incorrectly.

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the series-line label at the specified split characters. When a
series-line label is split, the label is split on each occurrence of the specified split
characters.

Default FALSE. The series-line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the series-line label can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
series-line label. In that case, all of the specified split characters together are treated
as a single split character.
When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the series-
line label is split unconditionally at each occurrence of any of the specified split
characters. If the series-line label does not contain any of the specified characters,
then the label is not split.
"character-list"
one or more characters with no delimiter between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the following
option:
SERIESPLOT Statement 751

curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the series-line label or are
dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the series-line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the series-line label text.
TRUE
drops the split characters from the series-line label text.
FALSE
includes the split characters in the series-line label text. When
CURVELABELSPLIT=TRUE and
CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split character, up
to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a series-line label with the following
specifications:
• CURVELABELPOSITION=MAX
• CURVELABEL="Product*Group*A"
• CURVELABELSPLIT=TRUE
• CURVELABELSPLITCHARDROP=TRUE | FALSE
• CURVELABELSPLITCHAR="*"
Note: The horizontal line to the left of the label represents the maximum end of the
series line for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the series-line label.
752 Chapter 6 • Plot Statements

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the series-line label block.
AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown in
the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which CURVELABELPOSITION=MAX.
Note: The horizontal line to the left of each label represents the maximum end of the
series line for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATALABEL=column | expression
specifies a column that supplies values for the data point labels.

Default No data labels are displayed

Note The label positions are adjusted to prevent the labels from overlapping.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.

Defaults For non-grouped data, the GraphDataText style element.

SERIESPLOT Statement 753

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the DATALABEL= option must also be
specified.

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

Note When the DATALABELPOSITION=AUTO option is in effect, in

some cases, the data label font size might be reduced in order to avoid
overlapping labels and markers.

Tip When the GROUP= option is used, each distinct group value might be
represented by a different color. The data label that is associated with
the group is assigned the group color. This option can be used to
specify a single color for all data labels in a plot, without affecting
items that have the group color, such as error bars and marker
symbols.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |

CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the vertices of the series line and
the markers, when displayed.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
data label. In that case, all of the specified split characters together are treated as a
single split character.
When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is
split unconditionally at each occurrence of any of the specified split characters. If the
data label does not contain any of the specified characters, then the label is not split.
754 Chapter 6 • Plot Statements

"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
TRUE
drops the split characters from the data label.
FALSE
includes the split characters in the data label. When DATALABELSPLIT=TRUE
and DATALABELSPLITCHARDROP=FALSE, each split character remains as
the last character in the current line. The characters that follow the split character,
up to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a data label with the following
specifications:
• the data label text for this label is Product*Group*A
• DATALABELSPLIT=TRUE
• DATALABELSPLITCHARDROP=TRUE | FALSE
• DATALABELSPLITCHAR="*"
SERIESPLOT Statement 755

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from the

label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains as the
last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.
AUTO
justifies the labels based on the DATALABELPOSITION= option, as shown in
the following table.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which DATALABELPOSITION=TOP.
Note: The gray vertical line at the bottom of each label represents the horizontal
center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of text.

The text box is anchored the same way that the unsplit text is anchored. For example,
if DATALABELPOSITION=TOP, then the bottom center of the text box is
positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

756 Chapter 6 • Plot Statements

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the series lines. The following figure shows a
grouped series plot with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the GraphSkins:DataSkin
style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the LINEATTRS= color.

DATATRANSPARENCY=number
specifies the degree of the transparency of the line, markers, line label, and data
labels, when displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISCRETEOFFSET=number
specifies an amount to offset all series lines and markers from discrete X values, or
discrete Y values, or both. This option is useful when graphing multiple response
variables side by side on a common axis

Default 0 (no offset, all series lines and markers are centered on the discrete X
values, or discrete Y values, or both)

Range -0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right on discrete X values and up on
discrete Y values. If the layout’s axis options set REVERSE=TRUE,
then the offset direction is also reversed.
SERIESPLOT Statement 757

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DISPLAY=STANDARD | ALL | display-options

specifies additional feature to display with the series line.
STANDARD
displays a series line without markers.
ALL
displays a series line with markers.
(display-options)
a space-separated list of one or more options enclosed in parentheses. Currently,
only the MARKERS option is supported, which displays a series line with
markers.

Default STANDARD

Tip Use the MARKERATTRS= and LINEATTRS= options to control the

appearance of the line and markers.

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline. When this option is TRUE, the fill color and outline color for
filled markers are determined in the following ways:
• If the GROUP= option is specified, then by default, the fill color is derived
from the GraphData1–GraphDataN style elements Color attribute, and the
marker outlined color is derived from the GraphData1–GraphDataN style
elements ContrastColor attribute.
• If the GROUP= option is not specified, then the marker fill is drawn by using
the MARKERFILLATTRS= specification, and the outline is drawn by using
the MARKEROUTLINEATTRS= specification.
FALSE
draws the markers using fill or an outline, but not both.

Default FALSE

Tip To specify the marker fill and outline colors for a non-grouped plot, set this
option to TRUE, and then use the MARKERFILLATTRS= and
MARKEROUTLINEATTRS= options to specify the colors.

See GROUP= on page 758

MARKERFILLATTRS= on page 766

MARKEROUTLINEATTRS= on page 766

758 Chapter 6 • Plot Statements

“boolean ” on page 1339 for other Boolean values that you can use.

GROUP=column | discrete-attr-var | expression

creates a separate series plot for each unique group value in the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color, line pattern, and marker symbol. Lines
and markers vary according to the ContrastColor, LineStyle, and
MarkerSymbol attributes of the GraphData1–GraphDataN and
GraphMissing style elements. Line thickness (for grouped and
ungrouped data) is controlled by the LINEATTRS= option.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of marker
symbols, colors, and line patterns.

The marker size is set by the MARKERATTRS= option.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

When both the GROUP= and the COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.

Tips The representations that are used to identify the groups can be
overridden. For example, each distinct group value is often
represented by a different line pattern, but you can use the
LINEATTRS=(PATTERN=pattern) option to assign the same line
pattern to all of the plot’s line patterns, letting line color indicate group
values. Likewise, you could use LINEATTRS=(COLOR= color) to
assign the same color to all lines, letting line pattern indicate group
values.

Starting with the second maintenance release of SAS 9.4, you can use
the LINECOLORGROUP=, LINEPATTERNGROUP=,
MARKERCOLORGROUP=, and MARKERSYMBOL= options to
assign line colors, line patterns, marker colors, and marker symbols
based on a different group column.

See “DISCRETEATTRVAR Statement” on page 1297

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.
OVERLAY
draws markers for a given group value at the exact coordinate. Depending on the
data, markers at a given coordinate might overlap.
SERIESPLOT Statement 759

CLUSTER
draws markers for a given group value adjacent to each other.

Interaction When markers are displayed and GROUPDISPLAY=CLUSTER is

in effect for interval data, the size of the markers in each cluster
might be reduced to no less than 5 pixels in order to display the
cluster within the smallest effective midpoint space. If you need
larger markers in that case, use the MARKERATTRS= option to
specify a larger marker size.

Tip Use the CLUSTERWIDTH= option to control the width of the

clusters when CLUSTER is in effect.

Default OVERLAY

Interaction For this option to take effect, the GROUP= option must also be
specified.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
DESCENDING
orders the groups within a category in descending order.

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.

Tips Use the CLUSTERWIDTH= option to control the distance between

the group markers in a cluster.
760 Chapter 6 • Plot Statements

Use the INDEX= option to alter the default sequence of visual

attributes that is assigned to the groups.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line pattern)
to one of the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y-column label. If a label is not defined, then the Y-column name
is used.
SERIESPLOT Statement 761

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the series line.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements, and the
GraphDataDefault:LineThickness style reference.

Interactions This option’s COLOR= suboption overrides the default behavior for
grouped data when the LINECOLORGROUP= option and the
COLORRESPONSE= option are not specified. When the COLOR=
suboption is specified without the LINECOLORGROUP= and
COLORRESPONSE= options, all lines have the same color.

The LINECOLORGROUP= option and the COLORRESPONSE=

option override this option’s COLOR= suboption.

This option’s PATTERN= suboption overrides the default behavior for

grouped data when the LINEPATTERNGROUP= option is not
specified. When the PATTERN= suboption is specified without the
LINEPATTERNGROUP= option, all lines have the same pattern.

The LINEPATTERNGROUP= option overrides this option’s

PATTERN= suboption.

See “General Syntax for Attribute Options” on page 1347 for the syntax
for using a style-element.

“Line Options” on page 1349 for available line-options.

LINECOLORGROUP=column | discrete-attr-var | expression

specifies a column that determines the line colors for a grouped plot independently of
the GROUP= column.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Note A discrete attribute map variable is valid for this option starting
with the third maintenance release of SAS 9.4.

When this option is specified with the GROUP= option, the series line colors are
selected from the GraphData1–GraphDataN style elements based on the
LINECOLORGROUP= column instead of on the GROUP= column.
762 Chapter 6 • Plot Statements

Default The line colors are selected based on the GROUP= column.

Requirement The column value must be the same for all of the points that define a
series line.

Interactions The GROUP= option must be specified for this option to have any
effect.

This option overrides the COLOR= suboption of the LINEATTRS=

option.

The COLORRESPONSE= option overrides this option.

Tip Use the LINEATTRS= option to set the line thickness.

See GROUP= on page 758

LINEATTRS=

LINEPATTERNGROUP=column | discrete-attr-var | expression

specifies a column that determines the line patterns for a grouped plot independently
of the GROUP= column.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Note A discrete attribute map variable is valid for this option starting
with the third maintenance release of SAS 9.4.

When this option is specified with the GROUP= option, the series line patterns are
selected from the GraphData1–GraphDataN style elements based on the
LINECOLORGROUP= column instead of on the GROUP= column.

Default The line patterns are selected based on the GROUP= column.

Requirement The column value must be the same for all of the points that define a
series line.

Interactions The GROUP= option must be specified for this option to have any
effect.

This option overrides the PATTERN= suboption of the

LINEATTRS= option.

Tip Use the LINEATTRS= option to set the line thickness.

See GROUP= on page 758

LINEATTRS=
SERIESPLOT Statement 763

LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to determine
the line thickness. By default, this option determines the thickness of the line that
represents the maximum response column value.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default Ten times the thickness that is specified by the GraphDataDefault style
element LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The LINETHICKNESSMAXRESPONSE= option specifies the

response value at which this maximum line thickness is reached. The
line thickness for response values that exceed the
LINETHICKNESSMAXRESPONSE= value are set to the value that
is specified by this option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMIN= option to specify the minimum line

thickness.

See “dimension” on page 1340

“Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line thickness.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default The maximum value in the response column that is specified in the
LINETHICKNESSRESPONSE= option.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The thickness for all lines that exceed the maximum response value is
set to the value specified in the LINETHICKNESSMAX= option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to determine
the line thickness.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
764 Chapter 6 • Plot Statements

Default The thickness specified by the GraphDataDefault style element

LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMAX= option to specify the maximum

line thickness.

See “dimension” on page 1340

“Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

LINETHICKNESSRESPONSE=numeric-column | range-attr-var | expression

specifies a response column or range attribute variable that is used to map a line
thickness to each group value.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

Tip For each range in the attribute map, the RANGEALTCOLOR= or

RANGEALTCOLORMODEL= option in the RANGE statement
determines the marker colors.

Default The GraphDataDefault style element LineThickness attribute.

Interactions When the column values are all zero, all negative, or all missing, this
option is ignored. In these cases, the default line thickness is used for
all of the lines.

This option overrides suboption THICKNESS= in the LINEATTRS=

option.

Note The LINETHICKNESSRESPONSE= values are assumed to be

constant for each group value in a grouped plot and for the entire plot
in an ungrouped plot. If the LINETHICKNESSRESPONSE column
has multiple values for a single GROUP value or ungrouped plot,
unpredictable results might occur.

See “Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

MARKERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the data markers.
SERIESPLOT Statement 765

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes of

the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions If FILLEDOUTLINEDMARKERS=TRUE, then this option’s

COLOR= suboption is ignored. In that case, to specify the marker fill
color, use the MARKERFILLATTRS= option instead.

This option’s COLOR= suboption overrides the default behavior for

grouped data when the MARKERCOLORGROUP= option is not
specified. When the COLOR= suboption is specified without the
MARKERCOLORGROUP option, all markers have the same color,
and the marker symbol alone distinguishes the markers.

The MARKERCOLORGROUP= option and theCOLORRESPONSE=

option override this option’s COLOR= suboption.

This option’s SYMBOL= suboption overrides the default behavior for

grouped data when the MARKERSYMBOLGROUP= option is not
specified. When the SYMBOL= suboption is specified without the
MARKERSYMBOLGROUP= option, all markers have the same
symbol, and the symbol color alone distinguishes the markers.

The MARKERSYMBOLGROUP= option overrides this option’s

SYMBOL= suboption.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

This option is ignored if the DISPLAY= option disables the display of

the markers.

If the DATASKIN= option is in effect, then the data skin determines

the marker outlines. Any outline-related settings from the current ODS
style or from the marker attribute options are ignored.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Marker Options” on page 1350 for available marker-options.

MARKERCOLORGROUP=column | discrete-attr-var | expression

specifies a column that determines the marker colors for a grouped plot
independently of the GROUP= column.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.
766 Chapter 6 • Plot Statements

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Note A discrete attribute map variable is valid for this option starting
with the third maintenance release of SAS 9.4.

When this option is specified with the GROUP= option, the series line markers are
selected from the GraphData1–GraphDataN style elements based on the
MARKERCOLORGROUP= column instead of on the GROUP= column.

Default The line marker colors are selected based on the GROUP= column.

Requirement The column value must be the same for all of the points that define a
series line.

Interactions The DISPLAY= option must enable the display of the series line
markers for this option to have any effect.

The GROUP= option must be specified for this option to have any
effect.

This option overrides the COLOR= suboption of the

MARKERATTRS= option.

The COLORRESPONSE= option overrides this option.

Tip Use the MARKERATTRS= option to set the marker symbol size.

See GROUP= on page 758

MARKERATTRS= on page 764

MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.

Defaults For non-grouped data, the COLOR attribute of the GraphDataDefault

style element

For grouped data, the COLOR attribute of a GraphData1–GraphDataN

style element

Restriction The TRANSPARENCY= fill option is ignored. Use the

MARKERATTRS= option to set the marker transparency.

Interaction This option is in effect only when

FILLEDOUTLINEDMARKERS=TRUE and the DISPLAY= option
enables fill display.

Note When style-element is specified, only the style element’s COLOR

attribute is used.

See “General Syntax for Attribute Options” on page 1347

“Fill Options” on page 1348

MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.
SERIESPLOT Statement 767

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, the LineThickness attritube of the GraphOutlines

style element and the ContrastColor attribute of a GraphData1–
GraphDataN style element.

Restriction The line style of the marker outline is always solid.

Interaction This option is ignored when a data skin is applied by the current style
or by the DATASKIN= option. In the latter case, the outline is set by
the data skin.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR and LINETHICKNESS attributes are used.

See “General Syntax for Attribute Options” on page 1347

“Line Options” on page 1349

MARKERSYMBOLGROUP=column | discrete-attr-var | expression

specifies a column that determines the marker symbols for a grouped plot
independently of the GROUP= column.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Note A discrete attribute map variable is valid for this option starting
with the third maintenance release of SAS 9.4.

When this option is specified with the GROUP= option, the series line marker
symbols are selected from the GraphData1–GraphDataN style elements based on the
MARKERSYMBOLGROUP= column instead of on the GROUP= column.

Default The line marker symbols are selected based on the GROUP= column.

Requirement The column value must be the same for all of the points that define a
series line.

Interactions The GROUP= option must be specified for this option to have any
effect.

This option overrides the SYMBOL= suboption of the

MARKERATTRS= option.

The DISPLAY= option must enable the display of the series line
markers for this option to have any effect.

Tip Use the MARKERATTRS= option to set the marker symbol size.

See GROUP= on page 758

768 Chapter 6 • Plot Statements

MARKERATTRS= on page 764

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, CURVELABEL, DATALABEL, GROUP,
and INDEX.

SMOOTHCONNECT=TRUE | FALSE
specifies that a smoothed line passes through all vertices.

Default FALSE
SERIESPLOT Statement 769

Interaction Starting with the third maintenance release of SAS 9.4, this option is
ignored when SPLINETYPE=QUADRATICBEZIER is in effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

SPLINEPOINTS=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the
INTERVAL= axis option.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default 20

Interaction When this option is set to a non-default value, markers and data labels,
when displayed, are positioned at their original data points.

SPLINETYPE=NONE | QUADRATICBEZIER
specifies the type of spline interpolation that is used to draw the series line.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
The following figure shows an example of SPLINETYPE= values
QUADRATICBEZIER and NONE.
Figure 6.2 Spline Types QUADRATICBEZIER and NONE

Default NONE

Interaction The SMOOTHCONNECT= option is ignored when this option is set to

a value other than NONE.

Note Markers and data labels, when displayed, are positioned at their original
data points as shown in Figure 6.2 on page 769.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the series line.
If this option is used, then the information specified replaces all the information that
is displayed by default. You can specify roles for columns that do not contribute to
the series plot along with roles that do.
(role-list)
an ordered, space-separated list of unique SERIESPLOT roles and user-defined
roles. SERIESPLOT roles include X, Y, CURVELABEL, DATALABEL,
GROUP, and COLORRESPONSE.
770 Chapter 6 • Plot Statements

Define user-defined roles with the ROLENAME= option.

Notes CURVELABEL is considered a role only when it is assigned a

column of values. It is not considered a role and does not display data
tips when it is assigned a string.

Starting with the third maintenance release of SAS 9.4, the

COLORRESPONSE role is valid.

Example The following example displays data tips for the columns assigned to
the roles X and Y as well as the column Obs, which is not assigned to
any pre-defined SERIESPLOT role. The Obs column must first be
assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X, Y, DATALABEL, COLORRESPONSE, and
GROUP.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
SERIESPLOT Statement 771

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a point or a segment of the curve is
selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
series line segment that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable points or segments of the curve, you must

include an ODS GRAPHICS ON statement that has the IMAGEMAP
option specified, and you must write the output to the ODS HTML
destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Note When you select a portion of a segment that is not an endpoint, the
nearest segment endpoint is used.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding segment is selected.

The URL value can be the same for any X and Y pair, meaning that
the same action is taken when the segment for those X and Y pairs is
selected.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.
772 Chapter 6 • Plot Statements

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
The SERIESPLOT statement is typically used to show time-dependent data.

Examples

Example 1: Grouped Series Plot

The following graph was generated by the “Example Program” on page 772:

Example Program
proc template;
define statgraph seriesplot;
begingraph;
entrytitle "Tech Stock Trends";
layout overlay;
seriesplot x=date y=close / group=stock name="stocks";
discretelegend "stocks";
endlayout;
endgraph;
end;
Example 2: Series Plot with Line-Thickness Response and Arrowheads 773

proc sgrender data=sashelp.stocks template=seriesplot;

where date > "31dec1999"d;
run;

Example 2: Series Plot with Line-Thickness Response and

Arrowheads
Starting with the third maintenance release of SAS 9.4, you can specify a response
variable to control the thickness of the lines in your grouped series plot. You can also
add arrowheads to one or both ends of your grouped or ungrouped series lines. This
example creates a grouped series plot that shows the monthly closing price for IBM,
Intel, and Microsoft stock in 2001. The mean volume is computed for each stock and is
used to control the thickness of the series line for each stock. The minimum line
thickness is set to 2px, and the maximum line width is set to 7px. A barbed arrowhead,
scaled to 1.25, is positioned at the end of each series line. In order to position the
arrowheads properly, the data must be sorted in ascending order by date. The following
figure shows the output.

Example Program
/* Extract the 2001 data from Sashelp.Stocks
and convert Volume to millions. */
data stocks;
set sashelp.stocks(where=
(date between "1jan02"d and "31dec02"d));
volume = volume / 1000000;
format date MONNAME3. volume 6.2;
;

/* Compute the average volume for each stock */

proc means data=stocks noprint;
by stock notsorted;
var volume;
output out=meanvolume(keep=stock meanvolume) mean=meanvolume;
run;
774 Chapter 6 • Plot Statements

/* Merge the average volume data with the stock data */

data stocks;
merge stocks meanvolume;
by stock;
run;

/* Sort the data by Date */

proc sort data=stocks;
by date;
run;

/* Define the graph template */

proc template;
define statgraph seriesplot;
begingraph / drawspace=wallpercent;
entrytitle "Stock Trends in 2001";
entryfootnote "Line thickness reflects the average volume.";
layout overlay /
xaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot)
type=discrete label="Month")
yaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot));
seriesplot x=date y=close / group=stock name="stocks"
arrowheadposition=end arrowheadshape=barbed
arrowheadscale=1.25
linethicknessresponse=meanvolume
linethicknessmin=2px linethicknessmax=7px
curvelabel=meanvolume;
drawtext "Average Volume (Millions)" / x=94 y=90
width=80 widthunit=pixel;
discretelegend "stocks";
endlayout;
endgraph;
end;

/* Render the graph */

proc sgrender data=stocks template=seriesplot;
run;

STEPPLOT Statement
Displays a series of horizontal and vertical line segments that connect observations of input data.

Syntax
STEPPLOT X=column | expression
Y=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
STEPPLOT Statement 775

ARROWHEADPOSITION= NONE | START | END | BOTH | column

specifies a position for arrowheads.
ARROWHEADSCALE=positive-number | numeric-column | expression
specifies an arrowhead scale factor based on the thickness of the arrow line.
ARROWHEADSHAPE= OPEN | FILLED | BARBED | column
specifies a shape for arrowheads.
BREAK=TRUE | FALSE
determines whether the plot line should show breaks at occurrences of
missing values of the Y column.
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group
clusters as a fraction of the minimum interval between adjacent data values.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the step plot lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the step lines, markers, error bars,
step-line labels, and data labels, when displayed.
DISPLAY=STANDARD | ALL | (display-options)
specifies whether to display markers on the step line.
ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
ERRORLOWER=numeric-column | expression
specifies the values of the lower endpoints on the Y error bars.
ERRORUPPER=numeric-column | expression
specifies the values of the upper endpoints on the Y error bars.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line
pattern) to one of the GraphData1–GraphDataN style elements.
JUSTIFY=(LEFT | CENTER | RIGHT)
specifies the location of the data point relative to the step.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the step line connecting the data points.
LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line thickness.
LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSRESPONSE=numeric-column | range-attr-var | expression
specifies a response column or range attribute variable that is used to map a
line thickness to each group value.
MARKERATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the data markers.
MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.
MARKEROUTLINEATTRS=style-element | (line-options)
776 Chapter 6 • Plot Statements

specifies the appearance of the marker outlines.

Axes options
CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Connect options
CONNECTORDER=XVALUES | XAXIS
specifies how to connect the data points to form the step line.
JOIN=TRUE | FALSE
specifies whether the steps are connected.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
step line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Error bar options

ERRORBARATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the error bars that are associated with the data
points.

Label options
CURVELABEL="string" | column | expression
specifies a label for the step line.
CURVELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the step-line labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the step-line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the step-line labels relative to the step line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the step-line label at the specified split characters.
CURVELABELSPLITCHAR="character-list"
STEPPLOT Statement 777

specifies one or more characters on which the step-line label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the step-line label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the step-line label
block.
DATALABEL=column | expression
specifies a column that supplies values for the data point labels.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |
CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the data points and
markers, when displayed.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all step lines and markers from discrete X
values when graphing multiple response variables side by side on a common
axis.
GROUP=column | discrete-attr-var | expression
creates a distinct set of lines, markers, and data labels for each unique group
value of the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped step lines and markers are overlaid or clustered
around the category midpoints.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a step line segment is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.
778 Chapter 6 • Plot Statements

Required Arguments
X=column | expression
specifies the column of the X values.
Y=numeric-column | expression
specifies the numeric column of the Y values.

Optional Arguments
ARROWHEADPOSITION= NONE | START | END | BOTH | column
specifies a position for arrowheads.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

NONE no arrowheads are displayed

START an arrowhead is displayed at the starting point of each line
END an arrowhead is displayed at the ending point of each line.
BOTH an arrowhead is displayed at each end of each line.
column specifies a column that provides an arrowhead position for each group
value.

Default NONE

Restriction When you specify a columnand the GROUP= option is in effect, the
arrowhead position values are assumed to be constant for each group
value. If the column has multiple values for a single group value, then
only one of the arrowhead position values is used for that group.

See “Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

ARROWHEADSCALE=positive-number | numeric-column | expression

specifies an arrowhead scale factor based on the thickness of the arrow line.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default 1.0

Restriction The arrowhead scale values are assumed to be constant for each line. If
a column or expression provides multiple values for a line, only one of
the values is used.

Interaction This option is ignored when ARROWHEADPOSITION=NONE is in

effect.

Tip Use a factor greater than 1.0 to make a larger arrowhead.

See “Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

ARROWHEADSHAPE= OPEN | FILLED | BARBED | column

specifies a shape for arrowheads.
STEPPLOT Statement 779

Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
The following figure shows each of the arrowhead shapes.

Default OPEN

Restriction When you specify a column and the GROUP= option is in effect, the
arrowhead shape values are assumed to be constant for each group
value. If the column has multiple values for a single group value, only
one of the arrowhead shape values is used for that group.

Interaction This option is ignored when ARROWHEADPOSITION=NONE is in

effect.

See “Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

BREAK=TRUE | FALSE
determines whether the plot line should show breaks at occurrences of missing
values of the Y column.

Default FALSE

Note When this option is set to FALSE, missing values are skipped and the line
continues through the missing value and to the next point.

See “boolean ” on page 1339 for other Boolean values that you can use.

CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when GROUPDISPLAY=CLUSTER.

AUTO uses the discrete axis for clustering groups when only one axis is
discrete. Uses the X axis for clustering if both axes are discrete or
interval.
X|Y uses the X or Y axis for clustering groups.

Default AUTO

Interaction The GROUPDISPLAY= option must be set to CLUSTER for this

option to have any effect.

CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters as a
fraction of the minimum interval between adjacent data values.

Default 0.85

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.
780 Chapter 6 • Plot Statements

Interactions For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

When markers are displayed for interval data and

GROUPDISPLAY=CLUSTER and CLUSTERWIDTH= are in effect,
the size of the markers in each cluster might be reduced to no less than
5 pixels in order to display the cluster within the smallest effective
midpoint space. If you need larger markers in that case, use the
MARKERATTRS= option to specify a larger marker size.

CONNECTORDER=XVALUES | XAXIS
specifies how to connect the data points to form the step line.
XVALUES
connects data points in the data order of the X column.
XAXIS
connects data points sorted by their X values.

Tip When the input data for the step lines is not sorted by the X column, you
can use XAXIS to assure the expected connect order.

Default XVALUES

CURVELABEL="string" | column | expression

specifies a label for the step line.

Restrictions When the GROUP= option is specified, "string" and expression are
not valid. Use column in that case.

When the GROUP= option is not specified, column is not valid. Use
"string" or expression in that case.

The line label for missing values is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

See GROUP= on page 791

CURVELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the step-line labels. When the GROUP=
option is used, each distinct group value might be represented by a different color.
The series label that is associated with the group is assigned the group color. This
option can be used to specify a single color for all series labels in a plot, without
affecting items that have the group color, such as lines and marker symbols.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, text color is derived from the

GraphData1:ContrastColor–GraphDataN:ContrastColor style
references. The font is derived from the GraphValueText style
element.

Interactions For this option to take effect, the CURVELABEL= option must also
be used.
STEPPLOT Statement 781

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the step-line label relative to the plot area.
INSIDE
locates the labels inside the plot area
OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the STEPPLOT is used in multicell

layouts such as LATTICE, DATAPANEL, or DATALATTICE, where
axes are external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the step-line
labels appear. For more information, see “Location and Position of
Curve Labels” on page 185.

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the step-line labels relative to the step line.
AUTO
automatically positions the step-line label near the step boundary along unused
axes whenever possible (typically Y2 and X2) in order to avoid collision with
tick values.

Restriction This option is used only when

CURVELABELPOSITION=OUTSIDE.

MAX
forces the step-line label to appear near maximum step values (typically, upper
right)
MIN
forces the step-line label to appear near minimum step values (typically, lower
left)
START
forces the step-line label to appear near the beginning of the steps.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the step line spirals around.
782 Chapter 6 • Plot Statements

END
forces the step-line label to appear near the end of the steps.

Restriction This option is used only when

CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the step line spirals around.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION=

INSIDE is specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the step-line
label appears. For more information, see “Location and Position of
Curve Labels” on page 185.

Note When you specify TICKVALUELIST=, VIEWMAX=, or

VIEWMIN= in an axis statement, the data points that are used to
determine the position of the step-line label might fall outside of the
graph area. In that case, the step-line label might not be displayed or
might be positioned incorrectly.

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the step-line label at the specified split characters. When a
step-line label is split, the label is split on each occurrence of the specified split
characters.

Default FALSE. The step-line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the step-line label can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
step-line label. In that case, all of the specified split characters together are treated as
a single split character.
When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the step-
line label is split unconditionally at each occurrence of any of the specified split
characters. If the step-line label does not contain any of the specified characters, then
the label is not split.
STEPPLOT Statement 783

"character-list"
one or more characters with no delimiter between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the following
option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the step-line label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the step-line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the step-line label text.
TRUE
drops the split characters from the step-line label text.
FALSE
includes the split characters in the step-line label text. When
CURVELABELSPLIT=TRUE and
CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split character, up
to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a step-line label with the following
specifications:
• CURVELABELPOSITION=MAX
• CURVELABEL="Product*Group*A"
• CURVELABELSPLIT=TRUE
• CURVELABELSPLITCHARDROP=TRUE | FALSE
• CURVELABELSPLITCHAR="*"
Note: The horizontal line to the left of the label represents the maximum end of the
step line for reference.
784 Chapter 6 • Plot Statements

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the step-line label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the step-line label block.
AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown in
the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which CURVELABELPOSITION=MAX.
Note: The horizontal line to the left of each label represents the maximum end of the
step line for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.
STEPPLOT Statement 785

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATALABEL=column | expression
specifies a column that supplies values for the data point labels.

Note The label positions are adjusted to prevent the labels from overlapping.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels. When the GROUP= option is
used, each distinct group value might be represented by a different color. The data
label that is associated with the group is assigned the group color. This option can be
used to specify a single color for all data labels in a plot, without affecting items that
have the group color, such as error bars and marker symbols.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the DATALABEL= option must also be
specified.

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

Note When the DATALABELPOSITION=AUTO option is in effect, in

some cases, the data label font size might be reduced in order to avoid
overlapping labels and markers.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |

CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the data points and markers, when
displayed.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
786 Chapter 6 • Plot Statements

separate split character unless the specified characters appear consecutively in the
data label. In that case, all of the specified split characters together are treated as a
single split character.
When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is
split unconditionally at each occurrence of any of the specified split characters. If the
data label does not contain any of the specified characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
TRUE
drops the split characters from the data label.
FALSE
includes the split characters in the data label. When DATALABELSPLIT=TRUE
and DATALABELSPLITCHARDROP=FALSE, each split character remains as
the last character in the current line. The characters that follow the split character,
up to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a data label with the following
specifications:
• the data label text for this label is Product*Group*A
• DATALABELSPLIT=TRUE
• DATALABELSPLITCHARDROP=TRUE | FALSE
• DATALABELSPLITCHAR="*"
STEPPLOT Statement 787

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from the

label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains as the
last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.
AUTO
justifies the labels based on the DATALABELPOSITION= option, as shown in
the following table.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which DATALABELPOSITION=TOP.
Note: The gray vertical line at the bottom of each label represents the horizontal
center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of text.

The text box is anchored the same way that the unsplit text is anchored. For example,
if DATALABELPOSITION=TOP, the bottom center of the text box is positioned at
the top of the marker.
788 Chapter 6 • Plot Statements

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the step plot lines. The following figure shows a
step plot with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the GraphSkins:DataSkin
style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interaction This option overrides the BEGINGRAPH statement DATASKIN=

option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the step lines, markers, error bars, step-
line labels, and data labels, when displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISCRETEOFFSET=number
specifies an amount to offset all step lines and markers from discrete X values when
graphing multiple response variables side by side on a common axis.

Default 0 (no offset, all step lines and markers are centered on the discrete X
values)
STEPPLOT Statement 789

Range -0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right. If the layout’s axis options set
REVERSE=TRUE, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DISPLAY=STANDARD | ALL | (display-options)

specifies whether to display markers on the step line.
STANDARD
displays a step line without markers.
ALL
displays a step line with markers.
(display-options)
a space-separated list of one or more options enclosed in parentheses. Currently,
only the MARKERS option is supported, which displays a step line with
markers.

Default STANDARD

Tip Use the MARKERATTRS= and LINEATTRS= options to control the

appearance of the line and markers.

ERRORBARATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the error bars that are associated with the data points.

Defaults For non-grouped data, the GraphError style element.

For grouped data, the LineStyle and LineThickness attributes of the

GraphError style element and the ContrastColor attribute of the
GraphData1–GraphDataN style elements. The LineStyle does not apply
to the "serif" parts of the error bars.

Interaction For this option to take effect, error bars must be displayed by the
ERRORLOWER= or ERRORUPPER= options.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
790 Chapter 6 • Plot Statements

Defaults SERIF in the first maintenance release of SAS 9.4 and earlier releases.

Starting with the second maintanance release of SAS 9.4,

GraphError:CapStyle style reference. If attribute CapStyle is not defined
in the active style, then SERIF is the default value.

Tip The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

ERRORLOWER=numeric-column | expression
specifies the values of the lower endpoints on the Y error bars.

Default The lower segment of the error bars is not drawn.

Requirement The error bar values must be absolute data values, not data values
relative to the value of the bar.

Interaction If the GROUP= option is specified and GROUPDISPLAY= STACK,

then this option is ignored.

Tip You can use the ERRORBARATTRS= option to control the

appearance of the error bars.

ERRORUPPER=numeric-column | expression
specifies the values of the upper endpoints on the Y error bars.

Default The upper segment of the error bars is not drawn.

Requirement The error bar values must be absolute data values, not data values
relative to the value of the bar.

Interaction This option is ignored when the GROUP= option is specified.

Tip You can use the ERRORBARATTRS= option to control the

appearance of the error bars.

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline. When this option is TRUE, the fill color and outline color for
filled markers are determined in the following ways:
• If the GROUP= option is specified, then by default, the fill color is derived
from the GraphData1–GraphDataN style elements Color attribute, and the
marker outlined color is derived from the GraphData1–GraphDataN style
elements ContrastColor attribute.
• If the GROUP= option is not specified, then the marker fill is drawn by using
the MARKERFILLATTRS= specification, and the outline is drawn by using
the MARKEROUTLINEATTRS= specification.
FALSE
draws the markers using fill or an outline, but not both.

Default FALSE
STEPPLOT Statement 791

Tip To specify the marker fill and outline colors for a non-grouped plot, set this
option to TRUE, and then use the MARKERFILLATTRS= and
MARKEROUTLINEATTRS= options to specify the colors.

See GROUP= on page 791

MARKERFILLATTRS= on page 797

MARKEROUTLINEATTRS= on page 797

“boolean ” on page 1339 for other Boolean values that you can use.

GROUP=column | discrete-attr-var | expression

creates a distinct set of lines, markers, and data labels for each unique group value of
the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color, line pattern, and marker symbol. Lines
and markers vary according to the ContrastColor, LineStyle, and
MarkerSymbol attributes of the GraphData1–GraphDataN and
GraphMissing style elements. Line thickness (for grouped and
ungrouped data) is controlled by the LINEATTRS= option.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of marker
symbols, colors, and line patterns.

The marker size is set by the MARKERATTRS= option.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Tip The representations that are used to identify the groups can be
overridden. For example, each distinct group value is often
represented by a different line pattern, but the
LINEATTRS=(PATTERN=pattern) option could be used to assign the
same line pattern to all of the plot’s line patterns, letting line color
indicate group values. Likewise, LINEATTRS=(COLOR=color) could
be used to assign the same color to all lines, letting line pattern
indicate group values.

See “DISCRETEATTRVAR Statement” on page 1297

GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped step lines and markers are overlaid or clustered around the
category midpoints.
OVERLAY
centers the step lines and markers for matching category values on the midpoints.
The step lines in each set of group values are superimposed on each other.
792 Chapter 6 • Plot Statements

CLUSTER
clusters the step lines and markers for matching category values around the
midpoints on a discrete axis or around the intervals on an interval axis. Each
cluster of group values is centered at the midpoint for the category.

Default OVERLAY

Restriction For this option to take effect, the GROUP= option must also be
specified.

Interaction When markers are displayed and GROUPDISPLAY=CLUSTER is in

effect for interval data, the size of the markers in each cluster might be
reduced to no less than 5 pixels in order to display the cluster within
the smallest effective midpoint space. If you need larger markers in that
case, use the MARKERATTRS= option to specify a larger marker size.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
DATA
orders the groups within a category in the group-column data order.
REVERSEDATA
orders the groups within a category in the reverse group-column data order.
Note: This feature applies to the second maintenance release of SAS 9.4 and to
later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.
DESCENDING
orders the groups within a category in descending order.

Default DATA

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes Attributes such as color, symbol, and pattern are assigned to each
group in the DATA order by default, regardless of the
GROUPORDER= option setting.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the unformatted
values. For character data, the order is based on the formatted values.
The data order of the observations and the visual attributes that are
assigned to the group values remain unchanged.

Tip Use the INDEX= option to alter the default sequence of visual
attributes that is assigned to the groups.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.
STEPPLOT Statement 793

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line pattern)
to one of the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

JOIN=TRUE | FALSE
specifies whether the steps are connected.

Default TRUE

See “boolean ” on page 1339 for other Boolean values that you can use.

JUSTIFY=(LEFT | CENTER | RIGHT)

specifies the location of the data point relative to the step.
794 Chapter 6 • Plot Statements

Default LEFT

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y-column label. If a label is not defined, then the Y-column name
is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the step line connecting the data points.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements, and the
GraphDataDefault:LineThickness style reference.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to determine
the line thickness. By default, this option determines the thickness of the line that
represents the maximum response column value.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default Ten times the thickness that is specified by the GraphDataDefault style
element LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The LINETHICKNESSMAXRESPONSE= option specifies the

response value at which this maximum line thickness is reached. The
line thickness for response values that exceed the
LINETHICKNESSMAXRESPONSE= value are set to the value that
is specified by this option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
STEPPLOT Statement 795

LINETHICKNESSMAXRESPONSE= option values is less than 0.5

for a line, that line is not drawn.

Tip Use the LINETHICKNESSMIN= option to specify the minimum line

thickness.

See “dimension” on page 1340

“Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line thickness.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default The maximum value in the response column that is specified in the
LINETHICKNESSRESPONSE= option.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The thickness for all lines that exceed the maximum response value is
set to the value specified in the LINETHICKNESSMAX= option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to determine
the line thickness.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default The thickness specified by the GraphDataDefault style element

LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMAX= option to specify the maximum

line thickness.

See “dimension” on page 1340

“Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.
796 Chapter 6 • Plot Statements

LINETHICKNESSRESPONSE=numeric-column | range-attr-var | expression

specifies a response column or range attribute variable that is used to map a line
thickness to each group value.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

Tip For each range in the attribute map, the RANGEALTCOLOR= or

RANGEALTCOLORMODEL= option in the RANGE statement
determines the marker colors.

Default The GraphDataDefault style element LineThickness attribute.

Interactions When the column values are all zero, all negative, or all missing, this
option is ignored. In these cases, the default line thickness is used for
all of the lines.

This option overrides suboption THICKNESS= in the LINEATTRS=

option.

Note The LINETHICKNESSRESPONSE= values are assumed to be

constant for each group value in a grouped plot and for the entire plot
in an ungrouped plot. If the LINETHICKNESSRESPONSE column
has multiple values for a single GROUP value or ungrouped plot,
unpredictable results might occur.

See “Example 2: Series Plot with Line-Thickness Response and

Arrowheads” on page 773 for an example of how to use this option.

MARKERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes of

the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions If FILLEDOUTLINEDMARKERS=TRUE, then this option’s

COLOR= suboption is ignored. In that case, to specify the marker fill
color, use the MARKERFILLATTRS= option instead.

This option’s COLOR= suboption overrides the default behavior for

grouped data. When the COLOR= suboption is specified in that case,
all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior for

grouped data. When the SYMBOL= suboption is specified in that
STEPPLOT Statement 797

case, all markers have the same symbol, and the symbol color alone
distinguishes the markers.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

This option is ignored if the DISPLAY= option disables the display of

the markers.

If the DATASKIN= option is in effect, then the data skin determines

the marker outlines. Any outline-related settings from the current ODS
style or from the marker attribute options are ignored.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Marker Options” on page 1350 for available marker-options.

MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.

Defaults For non-grouped data, the COLOR attribute of the GraphDataDefault

style element

For grouped data, the COLOR attribute of a GraphData1–GraphDataN

style element

Restriction The TRANSPARENCY= fill option is ignored. Use the

MARKERATTRS= option to set the marker transparency.

Interaction This option is in effect only when

FILLEDOUTLINEDMARKERS=TRUE and the DISPLAY= option
enables fill display.

Note When style-element is specified, only the style element’s COLOR

attribute is used.

See “General Syntax for Attribute Options” on page 1347

“Fill Options” on page 1348

MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, the LineThickness attritube of the GraphOutlines

style element and the ContrastColor attribute of a GraphData1–
GraphDataN style element.

Restriction The line style of the marker outline is always solid.

798 Chapter 6 • Plot Statements

Interaction This option is ignored when a data skin is applied by the current style
or by the DATASKIN= option. In the latter case, the outline is set by
the data skin.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR and LINETHICKNESS attributes are used.

See “General Syntax for Attribute Options” on page 1347

“Line Options” on page 1349

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

STEPPLOT Statement 799

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, CURVELABEL, DATALABEL,
ERRORLOWER, ERRORUPPER, GROUP, and INDEX.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the step line.
If this option is used, then it replaces all of the information that is displayed by
default. Roles for columns that do not contribute to the step plot can be specified
along with roles that do.
(role-list)
an ordered, space-separated list of unique STEPPLOT and user-defined roles.
STEPPLOT roles include X , Y , CURVELABEL , DATALABEL ,
ERRORLOWER , ERRORUPPER , and GROUP .
User-defined roles are defined with the ROLENAME= option.

Note CURVELABEL is considered a role only when it is assigned a

column of values. It is not considered a role and does not display data
tips when assigned a string.

Example The following example displays data tips for the columns assigned to
the roles X and Y as well as the column Obs, which is not assigned to
any pre-defined STEPPLOT role. The Obs column must first be
assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X , Y , DATALABEL , ERRORLOWER ,
ERRORUPPER , and GROUP .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
800 Chapter 6 • Plot Statements

TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a step line segment is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
step line segment that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable segments, you must include an ODS

GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Note When selecting a portion of a segment that is not an endpoint, the

nearest segment endpoint is used.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding segment is selected.
Example: STEPPLOT Statement 801

The URL value can be the same for any X and Y pairs, meaning that
the same action is taken when the segment for those X and Y pairs
are selected.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
For character columns, the X-axis is always of TYPE=DISCRETE. For numeric
columns, both the X- and the Y-axis are of TYPE=LINEAR by default. You can change
the axis type for numeric axes with the XAXISOPTS= and YAXISOPTS= options of the
containing overlay layout.
By default, the STEPPLOT statement uses the X values in data order. You can use the
CONNECTORDER= option to change the order.

Example: STEPPLOT Statement

The following graph was generated by the “Example Program” on page 802:
802 Chapter 6 • Plot Statements

Example Program
Here is the program code for this example.
proc template;
define statgraph stepplot;
begingraph;
entrytitle "Kaplan-Meier Survival Plot";
layout overlay /
yaxisopts=(linearopts=(viewmin=0 viewmax=1));
stepplot x=Time y=Survival /
group=Stratum name="step";
scatterplot x=Time y=Censored / name="scat"
legendlabel="Censored" markerattrs=(symbol=plus);
discretelegend "step" / location=inside
halign=right valign=top across=1;
discretelegend "scat" /
location=inside halign=center valign=top;
endlayout;
endgraph;
end;
run;

data Study;
input Group : $10. Time Status @@;
label Time="Time (Days)";
datalines;

... [datalines shown below]

run;

ods graphics;
SURFACEPLOTPARM Statement 803

ods exclude all;

ods output survivalplot=plotdata;
proc lifetest data=Study plots=(survival);
time Time * Status(0);
survival;
strata Group;
run;
ods select all;

proc sgrender data=plotdata template=stepplot;

run;

Here are the data lines for the example program:

Low-Risk 2569 0 Low-Risk 2506 0 Low-Risk 2409 0
Low-Risk 2218 0 Low-Risk 1857 0 Low-Risk 1829 0
Low-Risk 1562 0 Low-Risk 1470 0 Low-Risk 1363 0
Low-Risk 1030 0 Low-Risk 860 0 Low-Risk 1258 0
Low-Risk 2246 0 Low-Risk 1870 0 Low-Risk 1799 0
Low-Risk 1709 0 Low-Risk 1674 0 Low-Risk 1568 0
Low-Risk 1527 0 Low-Risk 1324 0 Low-Risk 957 0
Low-Risk 932 0 Low-Risk 847 0 Low-Risk 848 0
Low-Risk 1850 0 Low-Risk 1843 0 Low-Risk 1535 0
Low-Risk 1447 0 Low-Risk 1384 0 Low-Risk 414 1
Low-Risk 2204 1 Low-Risk 1063 1 Low-Risk 481 1
Low-Risk 105 1 Low-Risk 641 1 Low-Risk 390 1
Low-Risk 288 1 Low-Risk 421 1 Low-Risk 79 1
Low-Risk 748 1 Low-Risk 486 1 Low-Risk 48 1
Low-Risk 272 1 Low-Risk 1074 1 Low-Risk 381 1
Low-Risk 10 1 Low-Risk 53 1 Low-Risk 80 1
Low-Risk 35 1 Low-Risk 248 1 Low-Risk 704 1
Low-Risk 211 1 Low-Risk 219 1 Low-Risk 606 1
High-Risk 2640 0 High-Risk 2430 0 High-Risk 2252 0
High-Risk 2140 0 High-Risk 2133 0 High-Risk 1238 0
High-Risk 1631 0 High-Risk 2024 0 High-Risk 1345 0
High-Risk 1136 0 High-Risk 845 0 High-Risk 422 1
High-Risk 162 1 High-Risk 84 1 High-Risk 100 1
High-Risk 2 1 High-Risk 47 1 High-Risk 242 1
High-Risk 456 1 High-Risk 268 1 High-Risk 318 1
High-Risk 32 1 High-Risk 467 1 High-Risk 47 1
High-Risk 390 1 High-Risk 183 1 High-Risk 105 1
High-Risk 115 1 High-Risk 164 1 High-Risk 93 1
High-Risk 120 1 High-Risk 80 1 High-Risk 677 1
High-Risk 64 1 High-Risk 168 1 High-Risk 74 1
High-Risk 16 1 High-Risk 157 1 High-Risk 625 1
High-Risk 48 1 High-Risk 273 1 High-Risk 63 1
High-Risk 76 1 High-Risk 113 1 High-Risk 363 1

SURFACEPLOTPARM Statement
Creates a three-dimensional surface representing a response variable evaluated over a grid of X and Y
values.
Restriction: The SURFACEPLOTPARM statement does not support data tips.
804 Chapter 6 • Plot Statements

Requirements: For surface plots, the input data should form an evenly spaced grid of horizontal
values (X and Y) and one or more vertical values (Z) for each combination.
The input data must be sorted by Y and X in order to obtain the correct lighting.

Syntax
SURFACEPLOTPARM X=numeric-column | expression
Y=numeric-column | expression
Z=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
SURFACECOLORGRADIENT= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
starting with the second maintenance release of SAS 9.4, specifies the
column or range attribute map variable to use to determine the surface colors.
DATATRANSPARENCY=number
specifies the degree of the transparency of the surface.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the color of the filled surface or the wire-frame mesh.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.
SURFACECOLORGRADIENT=numeric-column | range-attr-var
specifies the column or range attribute map variable that is used to determine
the surface colors (in the first maintenance release of SAS 9.4 and earlier
releases).
SURFACETYPE=FILLGRID | FILL | WIREFRAME
specifies how the surface is displayed.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the X coordinates of the grid.
SURFACEPLOTPARM Statement 805

Y=numeric-column | expression
specifies the Y coordinates of the grid.
Z=numeric-column | expression
specifies the height of response values.

Note The input data should be sorted by both 1) the Y column and 2) the X
column. The sort direction for Y should be ascending. The sort direction of X
be either ascending or descending.

Optional Arguments
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
SURFACECOLORGRADIENT= option.
style-element
specifies the name of a style element. The style element should contain these
style attributes:

STARTCOLOR specifies a color for the smallest data value of the

COLORRESPONSE= or
SURFACECOLORGRADIENT= column.
NEUTRALCOLOR specifies a color for the midpoint of the range of the
COLORRESPONSE= or
SURFACECOLORGRADIENT= column.
ENDCOLOR specifies a color for the highest data value of the
COLORRESPONSE= or
SURFACECOLORGRADIENT= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorRamp style element.

Interaction For this option to have any effect, the COLORRESPONSE= or

SURFACECOLORGRADIENT= option must also be used.

Tip The REVERSECOLORMODEL= option can be used to reverse the

start and end colors of the ramp assigned to the color model.

COLORRESPONSE=numeric-column | range-attr-var | expression

starting with the second maintenance release of SAS 9.4, specifies the column or
range attribute map variable to use to determine the surface colors.
Note: Starting with the second maintenance release of SAS 9.4, the
COLORRESPONSE= option replaces the SURFACECOLORGRADIENT=
option. The syntax and functionality are the same. The
SURFACECOLORGRADIENT= option is still honored, but the
COLORRESPONSE= option is preferred.
806 Chapter 6 • Plot Statements

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.
You can use this option to add a second response variable to an analysis. For an
example, see “Example: SURFACEPLOTPARM Statement” on page 810.

Interaction Suboption COLOR= in the FILLATTRS= option overrides this option

for the fill colors.

Tip To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

DATATRANSPARENCY=number
specifies the degree of the transparency of the surface.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the color of the filled surface or the wire-frame mesh.

Default The GraphDataDefault:Color style reference.

Interaction The COLORRESPONSE= or SURFACECOLORGRADIENT= option

is ignored if this option is specified.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Z-column label. If a label is not defined, then the Z-column name
is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.
SURFACEPLOTPARM Statement 807

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
OVERLAY3D layout contribute to a common axis.

Default FALSE

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either the
ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page 1339 for other Boolean values that you can use.

SURFACECOLORGRADIENT=numeric-column | range-attr-var
specifies the column or range attribute map variable that is used to determine the
surface colors (in the first maintenance release of SAS 9.4 and earlier releases).
Note: Starting with the second maintenance release of SAS 9.4, the
SURFACECOLORGRADIENT= option is deprecated and replaced with the
COLORRESPONSE= option. The syntax and functionality are the same. The
SURFACECOLORGRADIENT= option is still honored, but the
COLORRESPONSE= option is preferred.

See COLORRESPONSE= on page 805

SURFACETYPE=FILLGRID | FILL | WIREFRAME

specifies how the surface is displayed.
FILLGRID
specifies a filled surface with superimposed grid lines
FILL
specifies a filled surface without grid lines
WIREFRAME
specifies an unfilled surface with grid lines

Default FILLGRID

Details
The SURFACEPLOTPARM statement assumes that the Z response values have been
provided for a uniform X-Y grid. Missing Z values leave a “hole” in the surface. The
observations in the input data set should form an evenly spaced grid of horizontal (X and
Y) values and one vertical (Z) value for each of these combinations. The observations
should be in sorted order of Y and X to obtain an accurate plot.
808 Chapter 6 • Plot Statements

The G3GRID procedure (requires a SAS/GRAPH license) can be used to interpolate the
necessary values to produce a data set with nonmissing Z values for every combination
of X and Y. The G3GRID procedure can also smooth data with spline interpolations. For
further details, see the documentation for PROC G3GRID in the SAS/GRAPH:
Reference.
Using PROC G3GRID, the following code performs a Spline interpolation and generates
this figure:

data nums;
do i=1 to 30;
X=10*ranuni(33)-5;
Y=10*ranuni(35)-5;
Z=sin(sqrt(x*x+y*y));
output;
end;
run;
proc g3grid data=nums out=gridded;
grid y*x=z / spline
axis1=-5 to 5 by 0.1
axis2=-5 to 5 by 0.1;
run;
proc sort data=gridded; by y x; run;

proc template;
define statgraph g3grid_surface;
begingraph;
entrytitle "Spline Interpolation";
layout overlay3d;
surfaceplotparm x=x y=y z=z /
surfacetype=fill;
endlayout;
endgraph;
SURFACEPLOTPARM Statement 809

end;
run;

proc sgrender data=gridded template=g3grid_surface;

run;

The KDE procedure can produce an output data set of gridded X-Y values where the Z
value is computed to be a Kernel Density Estimate of the distribution of X and Y. For
further details, see the documentation for PROC KDE in the SAS/STAT user’s guide.
Using PROC KDE on the nums data generated in the previous example, the following
code computes a Kernel Density Estimate and generates this figure:

/* use the nums data generated in the previous example */

proc kde data=nums;
bivar x y / ngrid=100
out=binned(rename=(value1=X value2=Y));
run;
proc sort data=binned; by y x;
label x="X" y="Y";
run;

proc template;
define statgraph kde_surface;
begingraph;
entrytitle "Kernel Density Estimate";
layout overlay3d;
surfaceplotparm x=x y=y z=density /
surfacetype=fill;
endlayout;
endgraph;
end;
run;
810 Chapter 6 • Plot Statements

proc sgrender data=binned template=kde_surface;

run;

The SURFACEPLOTPARM does not support the data tips that are enabled by the
IMAGEMAP= option in the ODS GRAPHICS statement.

Example: SURFACEPLOTPARM Statement

The following graph was generated by the “Example Program” on page 810:

Example Program
Here is the code for this example. The COLORRESPONSE= option is valid starting
with the second maintenance release of SAS 9.4. For prior releases, use the
SURFACECOLORGRADIENT= option instead.
proc template;
define statgraph surfaceplotparm;
begingraph;
entrytitle "Surface Plot of Lake Bed";
layout overlay3d / cube=false;
surfaceplotparm x=length y=width z=depth /
reversecolormodel=true
colorresponse=depth
colormodel=twocoloraltramp;
endlayout;
endgraph;
end;

/* create gridded data for surface

TEXTPLOT Statement 811

* proc g3grid is a sas/graph procedure */

proc g3grid data=sashelp.lake out=gridded;
grid width*length = depth / naxis1=75 naxis2=75;
run;

proc sgrender data= gridded template=surfaceplotparm;

run;

TEXTPLOT Statement
Displays text values at specific X and Y locations in the graph.
Note: This feature applies to the second maintenance release of SAS 9.4 and to later
releases.
Tip: Use the TEXTPLOT statement, rather than the SCATTERPLOT statement with the
MARKERCHARACTER= option, when you want more control over the appearance
of the text. The TEXTPLOT statement enables you to rotate the text to any angle,
manage the text position, split the text into multiple lines, display a bounding box
around the text, add a back-light effect to the text, and so on.

Syntax
TEXTPLOT X=column | expression
Y=column | expression
TEXT=column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
BACKLIGHT=number | AUTO
specifies a back-light effect for the marker text.
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group
clusters as a fraction of the minimum interval between adjacent data values.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute map variable to use to determine the
text-marker text colors.
CONTRIBUTEOFFSETS=ALL | NONE | (axis-offset-list)
specifies whether space requirements for this plot contribute to the
calculation of the axis offsets.
DATATRANSPARENCY=number
specifies the degree of the transparency of the text.
DISPLAY=STANDARD | ALL | (display-options)
specifies the features to display.
FILLATTRS=style-element | (fill-options)
specifies the appearance of the filled areas of the text.
INDEX=positive-integer-column | expression
812 Chapter 6 • Plot Statements

specifies indices for mapping marker text color to one of the GraphData1–
GraphDataN style elements.
OUTLINEATTRS=style-element | style-element(line-options) | (line-options)
specifies the appearance of the text-marker outlines.
PAD=dimension | (pad-options)
specifies the amount of extra space to add inside the text-marker border.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font properties of the text-marker text.

Axes options
CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
text values.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Experimental options
OUTFILE=fileref | "filename"
specifies a file for storing information about the text bounding-box for each
text value in the column specified in the OUTID= option.
OUTID=column | expression
specifies a column that contains text values to write to the file specified in the
OUTFILE= option.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
TEXTPLOT Statement 813

specifies the distance to offset all text values from discrete X values, discrete
Y values, or both.
GROUP=column | discrete-attr-var | expression
creates a separate text value for each unique group value in the specified
column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a text value is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Text options
FORMAT=format
specifies a SAS format or a user-defined format for the TEXT= column.
POSITION=position-option
specifies the position of the text value with respect to the location of the data
point.
ROTATE=number | numeric-column
specifies the angle of rotation, in degrees, for the text values.
SIZEMAX=dimension
specifies the maximum font size for a text marker when a response variable is
used to size the text-marker font.
SIZEMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum font size for
text markers.
SIZEMIN=dimension
specifies the minimum font size for text markers when a response variable is
used to size the font for text values.
SIZERESPONSE=numeric-column | numeric-expression
specifies a response column that is used to determine the font size for each
text value.
SPLITCHAR="character-list"
specifies one or more characters on which the text-marker text can be split.
SPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the text-marker text.
SPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the lines of text in the text-marker text blocks.
SPLITPOLICY=NONE | SPLIT | SPLITALWAYS
specifies a policy for avoiding collisions among the text values.
SPLITWIDTH=AUTO | width-in-characters
specifies the maximum width of each split line.
STRIP=TRUE | FALSE
814 Chapter 6 • Plot Statements

specifies whether leading and trailing blanks should be stripped from the text-
marker text before it is displayed.
VCENTER=BBOX | BASELINE
specifies whether the text is vertically centered with respect to the text
bounding box or to the text baseline.

Required Arguments
X=column
specifies the column for the X values.
Y=column
specifies the column for the Y values.
TEXT=column
specifies the column for the text values that are to be used for the markers.

Optional Arguments
BACKLIGHT=number | AUTO
specifies a back-light effect for the marker text. The effect is applied only to the
marker text.
number
specifies the degree of the back-light effect.

Range 0–1, where 0 specifies no effect and 1 specifies maximum effect

AUTO
the system selects an appropriate level for the back-light effect. If the GROUP=
or COLORRESPONSE= option is in effect, BACKLIGHT=0.75. Otherwise,
BACKLIGHT=0.5.
The following figure shows the effect on the text of an outlined, filled text marker.

The back light is based on text color. For dark colors, a contrasting white back-light
effect is used. For lighter colors, a contrasting black back-light effect is used. The
following figure shows the two back-light types when BACKLIGHT=1.

Default 0 (no back-light effect)

Restriction Vector graphics output cannot be generated when the back-light effect
is applied. If you request vector graphics output and enable the back-
light effect, an image is generated instead.
TEXTPLOT Statement 815

Tip The BACKLIGHT= option is most effective when the text color has a
low level of contrast with the background or when the background is
cluttered.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to the
color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute map variable to use to determine the text-
marker text colors.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

Tip For each range in the attribute map, the RANGECOLOR= or

RANGECOLORMODEL= option in the RANGE statement
determines the bounding-box fill colors. The
RANGEALTCOLOR= or RANGEALTCOLORMODEL= option
determines the text colors and the bounding-box outline colors.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
816 Chapter 6 • Plot Statements

COLORMODEL= option. When a range attribute map variable is specified, the

colors that are defined in the associated range attribute map are used instead.
You can use this option to add a second response variable to an analysis.

Interactions When the GROUP= option is specified with the COLORRESPONSE=

option, the GROUP= option is ignored.

Prior to the third maintenance release of SAS 9.4, suboption COLOR=

in the TEXTATTRS=, FILLATTRS=, and OUTLINEATTRS= options
overrides the color attributes that are normally determined by this
option.

Starting with the third maintenance release of SAS 9.4, interactions

between this option and suboption COLOR= in the TEXTATTRS=,
FILLATTRS=, and OUTLINEATTRS= options depend on the
DISPLAY= option settings. See “Response Colors in a Text Plot” on
page 833.

Note The gradient in a continuous legend for this plot reflects the text
colors.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

Use the OUTLINEATTRS= option to set the text bounding-box color

to a fixed color.

Use the FILLATTRS= option to set the text bounding-box fill color to
a fixed color or to modify the fill transparency.

Prior to the third maintenance release of SAS 9.4, when fill is

displayed and the COLORRESPONSE= option is in effect, the text
color and the fill color are derived from the color gradient, which
makes the text unreadable. In that case, use the BACKLIGHT= option
to add a backlight effect to the text, or use the TEXTATTRS= or
FILLATTRS= option to specify a different text or fill color.

CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when GROUPDISPLAY=CLUSTER.

AUTO uses the discrete axis for clustering groups when only one axis is
discrete. Uses the X axis for clustering if both axes are discrete or
interval.
X|Y uses the X or Y axis for clustering groups.

Default AUTO

Interaction The GROUPDISPLAY= option must be set to CLUSTER for this

option to have any effect.

CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters as a
fraction of the minimum interval between adjacent data values.

Default 0.85
TEXTPLOT Statement 817

Range 0.1–1, where 0.1 is the narrowest possible width and 1 is the widest
width.

Interaction For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

CONTRIBUTEOFFSETS=ALL | NONE | (axis-offset-list)

specifies whether space requirements for this plot contribute to the calculation of the
axis offsets. This plot’s layout container queries each of its plots for a preferred
offset and includes all of the offsets in the axis offset calculations. If the
DATALABEL= or MARKERCHARACTER= option is specified for this plot, this
plot might request a preferred offset that prevents the clipping of any data labels or
marker character strings that appear at the ends of the axes. The requested offset is
based on the longest string. If the label or marker character lengths vary significantly,
the result is wasted space when the shorter strings appear near the ends of the axes.
In that case, you can use the CONTRIBUTEOFFSETS= option to modify or
eliminate this plot’s contribution to the offset calculations in order to reclaim that
space.
ALL
the space requirements for this plot are contributed to the axis offset calculations.
NONE
the space requirements for this plot are not contributed to the axis offset
calculations.
(axis-offset-list)
a space-delimited list of specific contributions that this plot makes to the axis
offset calculations. The list is one or more of the following values, enclosed in
parentheses:

XMAX the space requirements for this plot are contributed to the X-axis
offset calculation for the maximum end.
XMIN the space requirements for this plot are contributed to the X-axis
offset calculation for the minimum end.
YMAX the space requirements for this plot are contributed to the Y-axis
offset calculation for the maximum end.
YMIN the space requirements for this plot are contributed to the Y-axis
offset calculation for the minimum end.

Default ALL

Interaction Offsets that are set in the layout axis options are always honored,
regardless of the setting on this option.

Note This option does not affect offset requests from other plots.

DATATRANSPARENCY=number
specifies the degree of the transparency of the text.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

818 Chapter 6 • Plot Statements

DISCRETEOFFSET=number
specifies the distance to offset all text values from discrete X values, discrete Y
values, or both.

Default 0 (no offset, all text values are centered on the discrete X values, or
discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
tick marks. A positive offset is to the right on discrete X values and is
up on discrete Y values. If the layout’s axis options set
REVERSE=TRUE, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DISPLAY=STANDARD | ALL | (display-options)

specifies the features to display.
STANDARD
displays the text only.
ALL
displays filled, outlined bounding boxes around the text.
(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:
FILL
displays filled bounding boxes around the text.

Tip Use the FILLATTRS= option to modify the appearance of the

bounding-box fill.

OUTLINE
displays outlined bounding boxes around the text.

Tip Use the OUTLINEATTRS= option to modify the appearance of the

bounding-box outline.

Default STANDARD

Restriction Vector graphics output cannot be generated when FILL or OUTLINE is

displayed. If you request vector graphics output and specify
DISPLAY=FILL or DISPLAY=OUTLINE, an image is generated
instead.

Tip When fill is displayed and the COLORRESPONSE= option is in effect,

a low contrast between the fill color and the text color can make some
of the text difficult to read or unreadable. In that case, use the
TRANSPARENCY= suboption to adjust the fill transparency or use the
BACKLIGHT= option to add a backlight effect to the text.
TEXTPLOT Statement 819

FILLATTRS=style-element | (fill-options)
specifies the appearance of the filled areas of the text. When fill options are
specified, only the COLOR= and TRANSPARENCY= suboptions are honored.

Defaults For non-grouped data, the Color attribute of the GraphDataDefault

style element.

For grouped data, the Color attribute of the GraphData1–GraphDataN

style elements.

Interactions For this option to have any effect, the fill must be enabled by the ODS
style or by the DISPLAY= option.

When this option’s COLOR= suboption is specified with the

GROUP= or COLORRESPONSE= option, the bounding-box fill color
is set to the COLOR= specification for all of the text values.

Tip The DATATRANSPARENCY= option sets the transparency for the

text-marker text, fill, and outlines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the text and
outlines but a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax
for using a style-element.

“Fill Options” on page 1348 for available fill-options.

FORMAT=format
specifies a SAS format or a user-defined format for the TEXT= column.

Default The format that is in effect for the column that is specified in the TEXT=
argument. If no format is in effect, BEST6 is used for numeric columns.

Note Not all of the SAS formats are supported. See Appendix 4, “SAS Formats
Not Supported,” on page 1353.

GROUP=column | discrete-attr-var | expression

creates a separate text value for each unique group value in the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Tip For each value in the discrete attribute map, the VALUE statement
TEXTATTRS= option determines the text color, the LINEATTRS=
option determines the bounding-box outline color, and the
FILLATTRS= option determines the bounding-box fill color and
transparency.

Interactions The group values are mapped in the order in which they appear in the
data.
820 Chapter 6 • Plot Statements

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

The COLORRESPONSE= option overrides the group settings for the

text color of the text value. In that case, text color is set according to
the gradient.

The COLOR= suboption of the TEXTATTRS= option overrides the

group settings for the text color of the text value. In that case, the text
color for all of the text values is set to the COLOR= specification.

Notes The legend entries for this plot reflect the text colors.

By default, for each text value, the bounding-box outline color is set to
the text color.

Tip Use the OUTLINEATTRS= option to set the text bounding-box color
to a fixed color.

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.
OVERLAY
draws text values for a given group value at the exact coordinate. Depending on
the data, markers at a given coordinate might overlap.
CLUSTER
draws text values for a given group value adjacent to each other.

Restriction CLUSTER is supported only when at least one axis is discrete.

Default OVERLAY

Tip Use the CLUSTERWIDTH= option to control the width of the clusters
when CLUSTER is in effect.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping marker text color to one of the GraphData1–
GraphDataN style elements.
TEXTPLOT Statement 821

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
a modulo operation remaps that index value to a number less than N
to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The TEXT-column label. If a label is not defined, then the TEXT-
column name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

OUTFILE=fileref | "filename"
specifies a file for storing information about the text bounding-box for each text
value in the column specified in the OUTID= option. The information is written in
the comma-separated value (CSV) format.
CAUTION:
OUTFILE= is an experimental option that is available in the third
maintenance release of SAS 9.4. Do not use this option in production jobs.

Interaction The OUTID= option must be specified for this option to have any
effect.
822 Chapter 6 • Plot Statements

Note This option is a specialized feature for users who want to customize the
placement of the text in a text plot.

Tip Use the IMPORT procedure to import the CSV values into a SAS data
set.

See “Customizing Text Marker Placement (Experimental)” on page 834

OUTID=column | expression
specifies a column that contains text values to write to the file specified in the
OUTFILE= option.
CAUTION:
OUTID= is an experimental option that is available in the third maintenance
release of SAS 9.4. Do not use this option in production jobs.

Interaction The OUTFILE= option must be specified for this option to have any
effect.

Note This option is a specialized feature for users who want to customize the
placement of the text in a text plot.

See “Customizing Text Marker Placement (Experimental)” on page 834

OUTLINEATTRS=style-element | style-element(line-options) | (line-options)

specifies the appearance of the text-marker outlines.

Defaults For non-grouped data, the ContrastColor attribute of the

GraphOutlines style element.

For grouped data, text values use the ContrastColor attribute of the
GraphData1–GraphDataN style elements. If the COLORRESPONSE=
option is specified, the outline colors vary according to the color
gradient.

Restriction This option uses only the color specification in the style element or
line options. The line pattern and line thickness specifications are
ignored.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or by the DISPLAY= option.

When the COLOR= suboption is specified with the GROUP= option

or with the COLORRESPONSE= option, for all of the text values, the
bounding-box border color is set to the COLOR= specification.

Note This plot’s legend entries reflect the marker text color.

See “General Syntax for Attribute Options” on page 1347 for the syntax
for using a style-element.

COLOR= on page 1349

PAD=dimension | (pad-options)
specifies the amount of extra space to add inside the text-marker border.
TEXTPLOT Statement 823

dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the text-marker border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options,
enclosed in parentheses:

LEFT=dimension TOP=dimension
RIGHT=dimension BOTTOM=dimension

Default Padding is a fraction of the font height.

Note Sides that are not assigned padding are padded with the default amount.

Tips This option is meaningful only when the DISPLAY= option displays
fills, outlines, or both.

Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page 1340

POSITION=position-option
specifies the position of the text value with respect to the location of the data point.
Specify one of the following position options:

BOTTOM CENTER TOP

BOTTOMLEFT LEFT TOPLEFT
BOTTOMRIGHT RIGHT TOPRIGHT
The VCENTER= option specifies whether the position is relative to the text bound
box or the text baseline. See the VCENTER= option. By default, the positions are
relative to the text bounding box. The following figure shows the effect of each of
these values on the position of an outlined text value when VCENTER=BBOX is in
effect. The red dot indicates the marker data-point location.

When CENTER, LEFT, or RIGHT is specified, and VCENTER=BASELINE is in

effect, the positions are relative to the text baseline, as shown in the following figure.
824 Chapter 6 • Plot Statements

Default CENTER

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either the
ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page 1339 for other Boolean values that you can use.

ROTATE=number | numeric-column
specifies the angle of rotation, in degrees, for the text values. Positive angles are
measured in a counter-clockwise direction, and negative angles are measured in a
clockwise direction. You can use an angle that exceeds 360 degrees in absolute
value.

Default 0

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)
TEXTPLOT Statement 825

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, GROUP, and COLORRESPONSE.

SIZEMAX=dimension
specifies the maximum font size for a text marker when a response variable is used
to size the text-marker font. By default, the font size of the text values that are
associated with the maximum response column value is set to the value specified by
this option.

Default Three times the size specified in the GraphDataText style element for
the maximum response column value marker.

Interactions The SIZERESPONSE= option must be specified for this option to

have any effect.

The SIZEMAXRESPONSE= option specifies the response value at

which the maximum font size for a text marker is reached. The font
size for all text values that exceed the SIZEMAXRESPONSE= value
is set to the value specified in this option.

Tips Use the SIZEMAXRESPONSE= option to specify the response value

at which the maximum font size for a text marker is reached.

Use the SIZEMIN= option to specify the minimum font size for text
markers.

See “dimension” on page 1340

SIZEMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum font size for text
markers.

Default The maximum value in the response column specified in the

SIZERESPONSE= option.

Interaction The SIZERESPONSE= option must be specified for this option to have
any effect.

Note The font size for all text values that exceed the maximum response
value is set to the value specified in the SIZEMAX= option.

SIZEMIN=dimension
specifies the minimum font size for text markers when a response variable is used to
size the font for text values.

Default The size specified in the GraphDataText style element for the minimum
response column value marker.

Interaction The SIZERESPONSE= option must be specified for this option to have
any effect.

Tip Use the SIZEMAX= option to specify the maximum text size.

See “dimension” on page 1340

826 Chapter 6 • Plot Statements

SIZERESPONSE=numeric-column | numeric-expression
specifies a response column that is used to determine the font size for each text
value.

Default The size specified in the GraphDataText style element for all text values.

Notes When the column value for an observation is 0, the font size for the text
value for that observation is set to the SIZEMIN= option value.

When the column value for an observation is negative or missing, the text
value for that observation is not displayed in the text plot. However, that
observation still contributes to the axis ranges, legend, and so on.

When all the column values are 0 or missing, this option is ignored. In that
case, the default font size is used for all of the text values.

Tip Use the SIZEMIN= and SIZEMAX= options to limit the minimum and
maximum font size for the text values.

SPLITCHAR="character-list"
specifies one or more characters on which the text-marker text can be split.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
data label. In that case, all of the specified split characters together are treated as a
single split character.
When SPLITPOLICY=SPLIT and a text value collision is detected, the text-marker
text is split on a specified split character only if a split is needed at that point in order
to make the text fit. In that case, a split might not occur on every split character.
When SPLITPOLICY=SPLITALWAYS, the text-marker text is split unconditionally
on every occurrence of a split character. If the text-marker text does not contain any
of the specified split characters, then the text is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
splitchar="abc"

The SPLITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interaction The SPLITCHARDROP= option specifies whether the split

characters are included in the displayed data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

TEXTPLOT Statement 827

SPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the text-marker text.
TRUE
drops a split character from the text-marker text when a split occurs at that
character. Split characters at which a split does not occur are left in place. The
SPLITPOLICY= option determines where the text is split. When
SPLITPOLICY=SPLIT, the text for each text value is split at a split character
only where a split is needed to fit the text in the available space. At each split
point, the split character is dropped, and the characters that follow the split
character, up to but not including the split character at the next split point, are
wrapped to the following line.
When SPLITPOLICY=SPLITALWAYS, the text-marker text is split at every
instance of a split character. All of the split characters are dropped. The
characters that follow each split character, up to but not including the next split
character, are wrapped to the next line.
FALSE
includes the split characters in the data label display. The SPLITPOLICY=
option determines how the split characters are displayed. When
SPLITPOLICY=SPLIT, each data label is split at a split character only where a
split is needed in order to make the label fit the available space. A split might not
occur at every split character in the label. At each split point, the split character
remains as the last character in the current line. The characters that follow the
split character, up to and including the split character at the next split point, are
then wrapped to the following line. This process repeats until all of the text is
displayed.
When SPLITPOLICY=SPLITALWAYS, the text for each marker is split at every
instance of a split character in the text regardless of whether a split is actually
needed. Each split character remains as the last character in the current line. The
characters that follow each split character, up to and including the next split
character, are then wrapped to the next line.

Default TRUE. A split character is dropped from the text-marker text when a
split occurs at that character.

Requirement The SPLITPOLICY= option must specify SPLIT or SPLITALWAYS.

See “boolean ” on page 1339 for other Boolean values that you can use.

SPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the lines of text in the text-marker text blocks.
AUTO
justifies the text based on the POSITION= option, as shown in the following
table.

POSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

828 Chapter 6 • Plot Statements

CENTER | LEFT | RIGHT

justifies the text center, left, or right, as specified.

Default AUTO

SPLITPOLICY=NONE | SPLIT | SPLITALWAYS

specifies a policy for avoiding collisions among the text values.
NONE
does not split the text for text values that collide.
SPLIT
splits the text-marker text at a split character only if a split is needed at that
character in order to make the text fit the available space. No split occurs at split
characters that occur where a split is not needed. If the text does not contain any
of the specified split characters, then a split does not occur. In that case, if the
text does not fit the available space, then it might collide with the adjoining text
values.

See the SPLITCHAR= option for information about specifying the split
characters

SPLITALWAYS
splits the text-marker text at every occurrence of a split character. If the text does
not contain any of the specified split characters, then a split does not occur.

See the SPLITCHAR= option for information about specifying the split
characters

Default NONE

SPLITWIDTH=AUTO | width-in-characters
specifies the maximum width of each split line.
AUTO
uses the width of the longest inter-split-character substring.
width-in-characters
specifies a fixed width, expressed as a character count.

Note When you specify a fixed width, the text-marker text is split
unconditionally every n characters, where n is the value of width-in-
characters.

Restriction This option has an effect only when SPLITPOLICY=SPLIT.

STRIP=TRUE | FALSE
specifies whether leading and trailing blanks should be stripped from the text-marker
text before it is displayed.

Default FALSE

Tip Stripping the blanks from the numeric value strings helps center each string
relative to its data point.

See “boolean ” on page 1339 for other Boolean values that you can use.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font properties of the text-marker text.
TEXTPLOT Statement 829

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the Font attribute of the GraphDataText style

element, and the ContrastColor attribute of a GraphDataN style
element.

Interactions When this option’s COLOR= suboption is used with the GROUP=
option, the COLOR= suboption specifies the color for all of the text
values.

This option’s COLOR= suboption overrides the COLORRESPONSE=

option. In that case, if a continuous legend is requested for the plot, the
legend is not drawn.

Note If one or more text options are specified and they do not include all the
font properties (color, family, size, weight, and style), the properties
that are not specified are derived from the GraphDataText style
element.

See “General Syntax for Attribute Options” on page 1347 for the syntax
for using a style-element.

“Text Options” on page 1351 for available text-options.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the text
values. If you use this option, it replaces all of the information that is displayed by
default. Roles for columns that do not contribute to the scatter plot can be specified
along with roles that do.
(role-list)
an ordered, space-separated list of unique TEXTPLOT roles and user-defined
roles. TEXTPLOT roles include X, Y, GROUP, and COLORRESPONSE.
User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns assigned to
the roles X and Y, as well as the column Obs, which is not assigned to
any pre-defined TEXTPLOT role. The Obs column must first be
assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to the following roles are automatically

included in the data tip information: X, Y, GROUP, and
COLORRESPONSE.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement with the IMAGEMAP option specified,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
830 Chapter 6 • Plot Statements

specified in the LINEAROPTS= or TIMEOPTS= option for either

axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a text value is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
text value that is an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable text values, you must include an ODS

GRAPHICS ON statement with the IMAGEMAP option specified,
and you must write the output to the ODS HTML destination.
TEXTPLOT Statement 831

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding point is selected.

The URL value can be the same for any X and Y pair. In that case,
the same action is taken when the points for that X and Y pair are
selected.

VCENTER=BBOX | BASELINE
specifies whether the text is vertically centered with respect to the text bounding box
or to the text baseline.
BBOX
vertically centers the text with respect to its bounding box.
BASELINE
vertically centers the text with respect to the text baseline. If the text is split into
multiple lines, the text is centered on the baseline of the last line of text.

Restriction This option is valid only when POSITION= is set to CENTER,

LEFT, or RIGHT. If POSITION= is set to any other value,
VCENTER=BBOX is used instead.

Default BBOX

Tip Use the POSITION= option to specify the text position with respect to the
text bounding box or to the text baseline.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.
832 Chapter 6 • Plot Statements

Details

About Text Plots

A text plot is basically a scatter plot that uses text as markers. To generate a text plot,
you can use the SCATTERPLOT statement with the MARKERCHARACTER= option
or you can use the TEXTPLOT statement. However, the TEXTPLOT statement provides
more control over the marker text. With the TEXTPLOT statement, the plot data
provides the marker location as X and Y coordinates, and the marker text. Each marker
consists of the marker text enclosed in a bounding box that is centered on its X and Y
coordinates. By default, the bounding box is not visible. You can specify options in the
TEXTPLOT statement to add an outline, background fill, or both, to the bounding box.
The following figure shows a simple example of a text marker with a filled and outlined
bounding box.

Using additional options in the TEXTPLOT statement, you can rotate the text around its
anchor point to any angle, manage the text position, split the text into multiple lines, add
a back-light effect to the text, and so on. If you currently use the SCATTERPLOT
statement and the MARKERCHARACTER= option to generate text plots, SAS
recommends that you use the TEXTPLOT statement instead.

Using Colors in a Text Plot

Text Marker Color Attributes

The TEXTPLOT statement enables you to control the following color attributes for the
text markers:
• text color
• bounding-box fill color
• bounding-box outline color
By default, only the marker text is displayed. Use the DISPLAY= option to specify
whether the bounding-box fill, bounding-box outline, or both are displayed.
The following table lists the options that you can use to control the color attributes.

Option Color Attribute Default

TEXTATTRS=(COLOR=color) text color For non-grouped data, the

GraphDataText:color style
reference.
For grouped data, the
GraphDataN:ContrastColor
style reference.
TEXTPLOT Statement 833

Option Color Attribute Default

FILLATTRS=(COLOR=color) bounding-box fill For non-grouped data, the

color GraphDataDefault:color style
reference.
For non-grouped data, the
GraphDataN:color style
reference.

OUTLINEATTRS=(COLOR=color) bounding-box outline For non-grouped data, the

color GraphOutlines:contrastColor
style reference.
For non-grouped data, the
GraphDataN:contrastColor
style reference.

When a response variable is used, in some cases, an attribute option is overridden by the
response color. See “Response Colors in a Text Plot” on page 833.

Group Colors in a Text Plot

The TEXTPLOT statement supports the GROUP= option, which enables you to create a
separate text marker for each unique value in a specified column. When the GROUP=
option is in effect, by default, the text marker color attributes are controlled by the
GraphData1–GraphDataN style elements. You can use the TEXTATTRS=,
FILLATTRS=, and OUTLINEATTRS= options to set one or more color attributes to a
fixed color. However, when you specify fixed colors, the contrast between the fixed
colors and the remaining variable colors might not be sufficient for some group values.
When both the GROUP= and COLORRESPONSE= options are in effect, the
COLORRESPONSE= option controls the color attributes.

Response Colors in a Text Plot

The TEXTPLOT statement supports the COLORRESPONSE= option, which enables
you to specify a numeric column or range attribute map variable that is used to
determine the text-marker colors. When a numeric column is specified, each unique
value is assigned a color from a color ramp that is specified in the COLORMODEL=
option. When a range attribute map variable is specified, each unique value is assigned
the attributes for that value that are defined in the attribute map. If the attribute map does
not define the color attributes for a value, that value is assigned a color from the color
ramp.
The TEXTPLOT statement COLORRESPONSE= option controls the color of the
marker text, fill, and outline, depending on which bounding-box attributes are displayed.
Prior to the third maintenance release of SAS 9.4, when the bounding-box fill and
outline are displayed, COLORRESPONSE= determines the color of the text and the
bounding-box fill. The bounding-box outline color in that case is determined by
suboption COLOR= in the OUTLINEATTRS= option. When only the outline is
displayed, the COLORRESPONSE= option determines the color of the text and the
bounding-box outline. You can use suboption COLOR= in the FILLATTRS=,
OUTLINEATTRS=, and TEXTATTRS= options to override the individual attributes.
Starting with the third maintenance release of SAS 9.4, when the bounding-box fill and
outline are displayed, the COLORRESPONSE= option determines the color of the text
and of the bounding-box outline. The bounding-box fill color in that case is determined
by suboption COLOR= in the FILLATRS= option. You can use suboption COLOR= in
the FILLATTRS=, OUTLINEATTRS=, and TEXTATTRS= options to override the
834 Chapter 6 • Plot Statements

individual attributes, depending on which bounding-box attributes are displayed. The

following table lists the results for some common color attribute option settings when the
COLORRESPONSE= option is in effect.

COLOR= Specified
in Attribute
Option: Result

TEXTATTRS= The result depends on the bounding-box display as follows:

• If the bounding-box fill, outline, or both are displayed, the text color
is set to its COLOR= specification. The bounding-box outline color
is determined by the GraphDataDefault style element, and the
bounding-box fill is assigned a color from the color ramp.
• If no bounding-box attributes are displayed, suboption COLOR= in
TEXTATTRS= is ignored, and the text is assigned a color from the
color ramp.

TEXTATTRS= and The result depends on the bounding-box display as follows:

FILLATTRS=
• If the fill and outline are displayed, the text and fill colors are set to
their COLOR= specification. The bounding-box outline is assigned
a color from the color ramp.
• If only the fill is displayed, the text color is set to its COLOR=
specification. Suboption COLOR= in FILLATTRS= is ignored, and
the bounding-box fill is assigned a color from the color ramp.

TEXTATTRS= and The result depends on the bounding-box display as follows:

OUTLINEATTRS=
• If the fill and outline are displayed, the text and outline colors are set
to their COLOR= specification. The bounding-box fill is assigned a
color from the color ramp.
• If only the outline is displayed, the outline color is set to its
COLOR= specification. Suboption COLOR= in TEXTATTRS= is
ignored, and the text is assigned a color from the color ramp.

TEXTATTRS=, When the bounding-box fill and outline are displayed, suboption
FILLATTRS=, and COLOR= in TEXTATTRS= is ignored, and the text is assigned a color
OUTLINEATTRS= from the color ramp. The fill and outline colors are set to their
COLOR= specification.

Customizing Text Marker Placement (Experimental)

Starting with the third maintenance release of SAS 9.4, you can use the TEXTPLOT
statement OUTFILE= and OUTID= options to write information about the bounding box
size and position for each text marker to a comma-separated values (CSV) file.
CAUTION:
OUTFILE= and OUTID= are experimental options that are available in the third
maintenance release of SAS 9.4. Do not use these options in production jobs.

These options are for users who want to customize the placement of the text in a text
plot. After you write the attributes to a CSV file, you can import the data into a SAS data
set, and then use the data to customize the placement of the text markers.
The following table lists the information that is written to the CSV file.
TEXTPLOT Statement 835

Column Name Data Type Description

DataDescent Numeric The descent, in y-axis units, of this observation's text

string. For a discrete axis, one unit is the distance
between adjacent discrete category ticks.

DataHeight Numeric The height, in y-axis units, of this observation's text

string. For a discrete axis, one unit is the distance
between adjacent discrete category ticks.

DataWidth Numeric The width, in x-axis units, of this observation's text

string. For a discrete axis, one unit is the distance
between adjacent discrete category ticks.

Descent Numeric The descent, in pixels, of this observation's text string.

Height Numeric The height, in pixels, of the bounding box for this
observation's text string.

Name Character The name that is specified in the NAME= option in this
TEXTPLOT statement. The name associates this CSV
file with a specific TEXTPLOT statement. This column
is useful when multiple CSV files from different
TEXTPLOT statements are merged into a single SAS
data set.

OutID Character The value for this observation from the column specified
in the OUTID= option in this TEXTPLOT statement.
The ID is used as a unique identifier for this
observation's text string.

RelativeDescent Numeric The relative descent of this observation's bounding box

as a proportion of the plot area width.

RelativeHeight Numeric The relative height of this observation's text string as a

proportion of the plot area height.

RelativeWidth Numeric The relative width of this observation's bounding box as

a proportion of the plot area width.

Width Numeric The width, in pixels, of the bounding box for this
observation's text string.

Here is an example of using the OUTFILE= and OUTID= options in a TEXTPLOT

statement to write text bounding-box information to CSV file textboxdata.csv. It also
shows how to import the CSV file into SAS data set Work.TextBoxData.
filename csvout "textboxdata.csv";

/* Write the text bound-box data to file textboxdata.csv */

proc template;
define statgraph textplot;
begingraph;
layout overlay;
textplot x=weight y=height text=name / name="textboxdata"
836 Chapter 6 • Plot Statements

outfile=csvout outid=name;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=textplot;

run;

/* Import the CSV file into data set Work.TextBoxData */

proc import
out=work.textboxdata
datafile=csvout
dbms=csv replace;
getnames=yes;
run;

Here is a partial listing of data set Work.TextBoxData.

The SAS System

Obs NAME OUTID WIDTH HEIGHT DESCENT RELATIVEWIDTH

1 textboxdata Alfred 34.045307 19.028384 5.004367 0.060364

2 textboxdata Alice 31.040453 19.028384 5.004367 0.055036
3 textboxdata Barbara 43.059872 19.028384 5.004367 0.076347
…
Obs RELATIVEHEIGHT RELATIVEDESCENT DATAWIDTH DATAHEIGHT DATADESCENT

1 0.046411 0.012206 6.489479 1.062499 0.279432

2 0.046411 0.012206 5.916715 1.062499 0.279432
3 0.046411 0.012206 8.207772 1.062499 0.279432
…

After you import the CSV file into a SAS data set, you can then use the data to position
the text values as needed.
VECTORPLOT Statement 837

Example: TEXTPLOT Statement

This example creates a text plot of weight by age and sex. Column Name provides the
text for filled, outlined markers. The following figure shows the output.

Here is the SAS code for this example.

proc template;
define statgraph textplot;
begingraph;
entrytitle "Weight by Age and Sex";
layout overlay / yaxisopts=(offsetmin=0.05 offsetmax=0.05);
textplot x=age y=weight text=name / name='textplot1'
display=all
textattrs=(weight=bold) fillattrs=(transparency=0.9)
group=sex groupdisplay=cluster clusterwidth=1;
discretelegend 'textplot1';
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.class template=textplot;

run;

VECTORPLOT Statement
Creates a plot of vectors (directed line segments).
838 Chapter 6 • Plot Statements

Syntax
VECTORPLOT X=numeric-column | expression
Y=numeric-column | expression
XORIGIN=numeric-constant | numeric-column | expression
YORIGIN=numeric-constant | numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the origin is considered when determining the data ranges
for the axes.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the line colors
to a continuous color gradient.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the vector plot lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the vector line and arrowhead, and
the vector labels.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the vector line and arrowhead.
LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line thickness.
LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSRESPONSE=numeric-column | range-attr-var | expression
specifies a response column or range attribute variable that is used to map a
line thickness to each group value.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

VECTORPLOT Statement 839

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
vector line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
specifies the labels at the ends of the vectors.
DATALABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |
CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the end points and arrow
heads.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of vector lines and data label colors for each unique
group value in the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Vector options
ARROWDIRECTION=OUT | IN |BOTH
specifies the placement of the arrowhead(s) at the end of the vector.
ARROWHEADS=TRUE | FALSE
specifies whether arrowheads are displayed on the vectors.
ARROWHEADSHAPE=OPEN | CLOSED | FILLED | BARBED
specifies the shape of the arrowheads.
840 Chapter 6 • Plot Statements

SCALE=number
specifies the scale factor of the vector length.

Required Arguments
X=numeric-column | expression
specifies the column for the X values of the vector endpoints.
Y=numeric-column | expression
specifies the column for the Y values of the vector endpoints.
XORIGIN=numeric-constant | numeric-column | expression
specifies the X data coordinate of the vector origin.
YORIGIN=numeric-constant | numeric-column | expression
specifies the Y data coordinate of the vector origin.

Optional Arguments
ARROWDIRECTION=OUT | IN |BOTH
specifies the placement of the arrowhead(s) at the end of the vector.
OUT
specifies a single arrowhead, pointing away from the origin, at the end of the
vector away from the origin.
IN
specifies a single arrowhead, pointing toward the origin, at the end of the vector
near the origin.
BOTH
specifies two arrowheads. One arrowhead points away from the origin, at the end
of the vector opposite from the origin. The other arrowhead points toward the
origin, at the end of the vector near the origin.

Default OUT

Interaction If ARROWHEADS= FALSE, then this option is ignored.

Tip Use the ARROWHEADSHAPE= option to control arrowhead

appearance.

ARROWHEADS=TRUE | FALSE
specifies whether arrowheads are displayed on the vectors.

Default TRUE

Interaction When this option is set to FALSE, the ARROWDIRECTION= and

ARROWHEADSHAPE= options are ignored and all vectors are
displayed as undirected line segments.

See “boolean ” on page 1339 for other Boolean values that you can use.

ARROWHEADSHAPE=OPEN | CLOSED | FILLED | BARBED

specifies the shape of the arrowheads.
VECTORPLOT Statement 841

Default OPEN

Interaction This option is ignored if ARROWHEADS= FALSE.

Note No arrowhead is drawn for a zero-length vector. A zero-length vector is

represented as a dot at its starting point.

Tip Use the ARROWDIRECTION= option to control arrow direction.

CLIP=TRUE | FALSE
specifies whether the origin is considered when determining the data ranges for the
axes.
FALSE
includes the origin when establishing the axis scales. Each axis might be
extended to force the display of the origin.
TRUE
ignores the origin when establishing axis scales. Each axis scale is determined by
the other plots in the overlay. This might result in the origin not being displayed
if its data range is not within the data ranges of tips of the vectors.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

842 Chapter 6 • Plot Statements

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the line colors to a
continuous color gradient.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column has
multiple values for a single GROUP value, unexpected results might
occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

Suboption COLOR= in the DATALABELATTRS= option overrides

this option for the data label color attribute.

This option overrides suboption COLOR= in the LINEATTRS=

option and varies the line color according to the color gradient or the
attribute map.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

For a numeric column or expression, the ThreeColorAltRamp style

element defines the line color gradient.

DATALABEL=column | expression
specifies the labels at the ends of the vectors.

Default No data labels are displayed

Note The label positions are automatically adjusted to prevent the labels from
colliding with other labels and other arrows.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.
VECTORPLOT Statement 843

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interaction For this option to take effect, the DATALABEL= option must also be
specified.

Note When the DATALABELPOSITION=AUTO option is in effect, in some

cases, the data label font size might be reduced in order to avoid
overlapping labels and markers.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT |

CENTER | RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the end points and arrow heads.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
data label. In that case, all of the specified split characters together are treated as a
single split character.
When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is
split unconditionally at each occurrence of any of the specified split characters. If the
data label does not contain any of the specified characters, then the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
844 Chapter 6 • Plot Statements

datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether

the split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
TRUE
drops the split characters from the data label.
FALSE
includes the split characters in the data label. When DATALABELSPLIT=TRUE
and DATALABELSPLITCHARDROP=FALSE, each split character remains as
the last character in the current line. The characters that follow the split character,
up to and including the next split character, are then wrapped to the next line.
The following figure shows an example of a data label with the following
specifications:
• the data label text for this label is Product*Group*A
• DATALABELSPLIT=TRUE
• DATALABELSPLITCHARDROP=TRUE | FALSE
• DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from the

label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains as the
last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.
VECTORPLOT Statement 845

See “boolean ” on page 1339 for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.
AUTO
justifies the labels based on the DATALABELPOSITION= option as shown in
the following table.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER | LEFT | RIGHT

justifies the labels center, left, or right, as specified.
The following figure shows an example in which DATALABELPOSITION=TOP.
Note: The gray vertical line at the bottom of each label represents the horizontal
center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of text.

The text box is anchored the same way that the unsplit text is anchored. For example,
if DATALABELPOSITION=TOP, then the bottom center of the text box is
positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the vector plot lines. The following figure shows a
vector plot with each of the skins applied.
846 Chapter 6 • Plot Statements

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the GraphSkins:DataSkin
style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interaction This option overrides the BEGINGRAPH statement DATASKIN=

option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the vector line and arrowhead, and the
vector labels.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

GROUP=column | discrete-attr-var | expression

creates a distinct set of vector lines and data label colors for each unique group value
in the specified column.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color and line pattern. Lines vary according
to the ContrastColor and LineStyle attributes of the GraphData1–
GraphDataN and GraphMissing style elements.
VECTORPLOT Statement 847

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line patterns
and colors.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

When both the GROUP= and the COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.

Tip You can use the LINEATTRS= option to override the representations
that are used to identify the groups. For example, You can use
LINEATTRS=(PATTERN=SOLID) to assign the same pattern to all of
the lines, letting the line color distinguish group values. Likewise, you
can use LINEATTRS=(COLOR=BLACK) to assign the same color to
all of the lines, letting the line pattern distinguish group values.

See “DISCRETEATTRVAR Statement” on page 1297

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in effect,
the INDEX= option is used, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page 1339 for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N, then
848 Chapter 6 • Plot Statements

a modulo operation remaps that index value to a number less than N

to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 183.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the vector line and arrowhead.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor, LineStyle, and LineThickness

attributes of the GraphData1–GraphDataN style elements.

Interactions The COLORRESPONSE= option overrides this option’s COLOR=

suboption.

The LINETHICKNESSRESPONSE= option overrides this option’s

THICKNESS= suboption.

Note The arrow head size is nonlinearly proportional to the line thickness in
order to maintain appropriately sized arrow heads for thicker lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element

“Line Options” on page 1349 for available line-options.

LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to determine
the line thickness. By default, this option determines the thickness of the line that
represents the maximum response column value.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default Ten times the thickness that is specified by the GraphDataDefault style
element LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The LINETHICKNESSMAXRESPONSE= option specifies the

response value at which this maximum line thickness is reached. The
line thickness for response values that exceed the
VECTORPLOT Statement 849

LINETHICKNESSMAXRESPONSE= value are set to the value that

is specified by this option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMIN= option to specify the minimum line

thickness.

See “dimension” on page 1340

LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line thickness.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default The maximum value in the response column that is specified in the
LINETHICKNESSRESPONSE= option.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The thickness for all lines that exceed the maximum response value is
set to the value specified in the LINETHICKNESSMAX= option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to determine
the line thickness.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default The thickness specified by the GraphDataDefault style element

LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMAX= option to specify the maximum

line thickness.

See “dimension” on page 1340

850 Chapter 6 • Plot Statements

LINETHICKNESSRESPONSE=numeric-column | range-attr-var | expression

specifies a response column or range attribute variable that is used to map a line
thickness to each group value.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

Tip For each range in the attribute map, the RANGEALTCOLOR= or

RANGEALTCOLORMODEL= option in the RANGE statement
determines the marker colors.

Default The GraphDataDefault style element LineThickness attribute.

Interactions When the column values are all zero, all negative, or all missing, this
option is ignored. In these cases, the default line thickness is used for
all of the lines.

This option overrides suboption THICKNESS= in the LINEATTRS=

option.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for determining
default axis features. This option is needed only when two or more plots within an
overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the primary
plot on a per-axis basis. When no plot is designated as the primary plot,
the first plot that can be a primary plot is considered the primary plot. If
multiple plots specify PRIMARY=TRUE for the same axis, then the
last such plot encountered is considered the primary plot.

See “When Plots Share Data and a Common Axis” on page 880

“boolean ” on page 1339 for other Boolean values that you can use.
VECTORPLOT Statement 851

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, DATALABEL, XORIGIN, YORIGIN,
GROUP, and INDEX.

SCALE=number
specifies the scale factor of the vector length.

Default 1.0

Restriction The number specified must be greater than 0.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a vector line.
If this option is used, then the information specified replaces all of the information
that is displayed by default. You can specify roles for columns that do not contribute
to the vector plot along with roles that do.
(role-list)
an ordered, space-separated list of unique VECTORPLOT roles and user-defined
roles. VECTORPLOT roles include X, Y, DATALABEL, XORIGIN, YORIGIN,
GROUP, INDEX, and COLORRESPONSE.
Define user-defined roles with the ROLENAME= option.

Note Starting with the third maintenance release of SAS 9.4, the
COLORRESPONSE role is valid.

Example The following example displays data tips for the columns assigned to
the roles X, Y, GROUP, and the column Obs, which is not assigned to
any pre-defined VECTORPLOT role. The Obs column must first be
assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y GROUP)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: X, Y, DATALABEL, XORIGIN, YORIGIN,
GROUP, and COLORRESPONSE.
852 Chapter 6 • Plot Statements

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X
Example: VECTORPLOT Statement 853

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
By default in a VECTORPLOT, each vector starts at 0, 0 in the data space and is
terminated with an open arrowhead. Zero-length vectors are represented by a dot at the
starting point. To specify alternative coordinates for the starting point, use the
XORIGIN= and YORIGIN= arguments.

Example: VECTORPLOT Statement

The following graph was generated by the “Example Program” on page 853:

Example Program
data CarPref;
input Make $11. Model $13. (S1-S25) (1.);
854 Chapter 6 • Plot Statements

datalines;
Cadillac Deville 8007990491240508971093809
Chevrolet Aveo 0051200423451043003515698
Chevrolet Cavalier 4053305814161643544747795
Chevrolet Malibu 6027400723121345545668658
Dodge Intrepid 7006000434101107333458708
Dodge Stratus 3005005635461302444675655
Dodge Neon 4005003614021602754476555
Ford Taurus 2024006715021443530648655
Ford Mustang 5007197705021101850657555
Ford Focus 0021000303030201500514078
Honda Accord 5956897609699952998975078
Honda Civic 4836709507488852567765075
Lincoln LS 7008990592230409962091909
Pontiac Firebird 0107895613201206958265907
Volkswagen Jetta 4858696508877795377895000
Volkswagen Beetle 4858509709695795487885000
Volvo S40 9989998909999987989919000
;
run;
* Compute Two Component Model;
ods graphics;
ods exclude all;
ods output mdprefplot=plotdata;
proc prinqual data=CarPref n=2 replace mdpref method=mgv;
id model;
transform monotone(S1-S25);
run;
ods select all;

proc template;
define statgraph vectorplot;
begingraph;
entrytitle "Multidimensional Preference Analysis";
entrytitle "of Preference Ratings for Automobiles";
layout overlayequated / equatetype=fit cycleattrs=true;
referenceline y=0 / datatransparency=0.7;
referenceline x=0 / datatransparency=0.7;
vectorplot y=vec2 x=vec1 xorigin=0 yorigin=0 /
datalabel=label2var;
scatterplot y=prin2 x=prin1 /
datalabel=idlab1 primary=true
markerattrs=(symbol=circlefilled);
endlayout;
endgraph;
end;
run;

proc sgrender data=plotdata template=vectorplot;

run;

WATERFALLCHART Statement
Creates a waterfall chart that is computed from input data.
WATERFALLCHART Statement 855

Interaction: A Waterfall chart accumulates response values in data order. Any change in the
order of the X-axis values from the data order can adversely affect the waterfall
chart. The X-axis value order can change when the Waterfall chart is overlaid with
other plots or when it is used in a Lattice with uniform axes. It can also change when
certain options are applied to the X axis.
Tip: Starting with the third maintenance release of SAS 9.4, you can use subpixel
rendering with this statement. It is enabled by default. To disable subpixel rendering,
specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 33. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” in SAS ODS Graphics:
Procedures Guide.

Syntax
WATERFALLCHART CATEGORY=column | expression
RESPONSE=numeric-column | expression </option(s)>;

Summary of Optional Arguments

Appearance options
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.
COLORGROUP=column | discrete-attr-var | expression
specifies a column that is used to discretely color the transaction bars.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the numeric column or range attribute map variable that is used to
determine the transaction-bar colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, bar labels,
and trend lines, if displayed.
DISPLAY=STANDARD | ALL | (display-options)
specifies which bar features to display.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled transaction bars.
FILLTYPE=SOLID | GRADIENT
determines whether a solid or gradient fill is used in the bars.
FINALBARATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the “final” bar, if displayed.
INITIALBARATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the “initial” bar, if displayed.
INITIALBARVALUE=number
specifies a value for the initial bar.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bar outlines.
856 Chapter 6 • Plot Statements

Axes options
BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline.
FINALBARTICKVALUE="string"
specifies a tick value to use on the category axis when the “final” bar is
displayed
INITIALBARTICKVALUE="string"
specifies a tick value to use on the category axis when the “initial” bar is
displayed.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
BARLABEL=TRUE | FALSE
specifies whether the bar statistic value is displayed at the end of the bar.
BARLABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the text properties of the bar label text.
BARLABELFITPOLICY=AUTO | NONE
specifies a policy for avoiding collisions among the bar labels when labels
are displayed.
BARLABELFORMAT=format
specifies the text format used to display the bar label.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

ODS options
URL=string-column
specifies an HTML page to display when a bar is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
COLORSTAT=SUM | MEAN
specifies the statistic to use for computing the response colors.
WATERFALLCHART Statement 857

STAT=SUM | MEAN
specifies the statistic to be computed for the RESPONSE axis.

Trend line options

CONNECT=START | END
determines whether trend lines connect to the adjacent bar’s starting or
ending value.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the trend lines that connect the bars.
CONNECTDECREASINGATTRS=style-element | style-element (line-options) |
(line-options)
specifies the appearance of trend lines that denote a decreasing value between
bars.
CONNECTINCREASINGATTRS=style-element | style-element (line-options) |
(line-options)
specifies the appearance of trend lines that denote an increasing value
between bars.

Required Arguments
CATEGORY=column | expression
specifies the column or expression for the category values. Duplicated category
values are summarized into a unique value. All values are treated as discrete.
RESPONSE=numeric-column | expression
specifies the numeric column or expression for the response values.

Optional Arguments
BARLABEL=TRUE | FALSE
specifies whether the bar statistic value is displayed at the end of the bar.

Default FALSE

Tip The font and color attributes for the label are specified by the
BARLABELATTRS= option. The text format is specified by the
BARLABELFORMAT= option.

See “boolean ” on page 1339 for other Boolean values that you can use.

BARLABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the bar label text.

Default The GraphDataText style element.

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

BARLABELFITPOLICY=AUTO | NONE
specifies a policy for avoiding collisions among the bar labels when labels are
displayed.
858 Chapter 6 • Plot Statements

AUTO
rotates the bar labels if the labels exceed the midpoint spacing. If the labels
collide horizontally due to thin bars, then AUTO drops all of the labels. The
following figure shows an example.

See the BARWIDTH= option for more information about the bar spacing.
NONE
does not rotate the bar labels. Labels that are too long overlap.

Default AUTO

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

BARLABELFORMAT=format
specifies the text format used to display the bar label.

Default The column format assigned to the RESPONSE= column or BEST6

if no format is assigned.

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.

Default 0.85

Range 0.1–1, where 0.1 is the narrowest and 1 is the widest

Notes This option is needed only to change the default behavior.

By default, the bar width automatically adjusts based on the number of bars
to be displayed and the wall width.

Tip To remove any inter-bar gap, set BARWIDTH=1.

BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.
WATERFALLCHART Statement 859

Default The GraphAxisLines style element.

Notes The baseline is always drawn by default.

When style-element is specified, only the style element’s COLOR,

LINESTYLE, and LINETHICKNESS attributes are used.

Tip To suppress the baseline, set the line thickness to 0:

baselineattrs=(thickness=0)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline. The baseline is always
displayed in the chart, whether for a specified value or for the default value. When
this option is used, the axis range is adjusted to include the baseline, and the baseline
is placed at the specified value on the response axis.

Default 0

Interactions The value set by this option affects only the chart’s initial and final
bars. If no initial bar value is specified, then the first transaction bar is
drawn from 0, no matter what is set for the baseline value.

If necessary, the response axis data range is extended to include the

baseline intercept. When a logarithmic response axis is requested and
BASELINEINTERCEPT= specifies 0 or a negative value, the
response axis reverts to a linear axis. To restore the log axis in that
case, set BASELINEINTERCEPT= to a positive value.
860 Chapter 6 • Plot Statements

Note Label positions are automatically adjusted to prevent the labels from
overlapping.

Tips Control the appearance of the baseline with the BASELINEATTRS=

option.

To suppress the baseline, use the BASELINEATTRS= option to set

the line thickness to 0.

COLORGROUP=column | discrete-attr-var | expression

specifies a column that is used to discretely color the transaction bars.
discrete-attr-var
specifies a discrete attribute map variable that is defined in a
DISCRETEATTRVAR statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Interactions This option is ignored if the COLORRESPONSE= option is also

specified.

If a column or expression is specified, then the unique column values

are found and the transaction bar colors are derived from the
GraphData1–GraphDataN style elements. The COLOR attribute is
used for the bar fill colors and the CONTRASTCOLOR attribute is
used for the bar outline colors.

If a variable that is associated with an attribute map is specified, then

the color mapping defined by the associated DISCRETEATTRMAP
statement is used for the transaction bars.

Notes All of the COLORGROUP values for a specific category value must
be the same. Otherwise, the results are unpredictable.

All the transaction bars have only one fill and one outline color,
determined by the ODS style or set by the FILLATTRS= and
OUTLINEATTRS= options.

Tip To manage the color of the “initial” bar, use the

INITIALBARATTRS= option. To manage the color of the “final” bar,
use the FINALBARATTRS= option.

See “DISCRETEATTRMAP Statement” on page 1287

“DISCRETEATTRVAR Statement” on page 1297

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
color-ramp-style-element
specifies the name of a color-ramp style element. The style element should
contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= column.
WATERFALLCHART Statement 861

NEUTRALCOLOR specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= column.
(color-list)
specifies a space-separated list of colors to use in the color ramp. You can use
style attribute references such as GraphData3:Color, color names, or RGB,
CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page 1340

Default The ThreeColorRamp style element

Interactions For this option to take effect, the COLORRESPONSE= option must
also be specified.

When FILLTYPE=GRADIENT and a color list is specified, the

middle color in the list is treated as the NEUTRAL color.

Tip To manage the color of the initial bar, use the INITIALBARATTRS=
option. To manage the color of the final bar, use the
FINALBARATTRS= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the numeric column or range attribute map variable that is used to
determine the transaction-bar colors.
range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Restriction This option affects only the fill colors. When only the bar outlines are
displayed, this option has no effect.

Interactions When the COLORGROUP= option is specified with this option, the
COLORGROUP= option is ignored.

This option overrides suboption COLOR= in the FILLATTRS=

option.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.
862 Chapter 6 • Plot Statements

Use the COLORSTAT= option to specify the statistic to compute for

the COLORRESPONSE= column.

To produce discrete color mapping, the RANGEATTRMAP statement

can define an attribute map that maps a single color to all values
greater than 0, and another color to all values less than 0.

Use the FILLTYPE= option to indicate whether the color mapping is

used to produce solid or gradient fills. When FILLTYPE=GRADIENT
is in effect, the color at the end of the bar is based on the color
mapping, and the neutral color of the color ramp is used as the starting
color of each bar.

COLORSTAT=SUM | MEAN
specifies the statistic to use for computing the response colors.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default SUM

Interaction This option is ignored when the COLORRESPONSE=option is not

specified.

Tip This option might affect existing SAS programs. For programs written
before the third maintenance release of SAS 9.4, if STAT= and
COLORRESPONSE= are specified in a WATERFALLCHART
statement, the chart colors and color statistic might change from the
previous SAS releases. To restore the original colors and color statistic
in that case, set COLORSTAT= in the WATERFALLCHART statement
to the same statistic that is specified in STAT=.

CONNECT=START | END
determines whether trend lines connect to the adjacent bar’s starting or ending value.
START
draws the trend lines horizontally and connects each to the adjacent bar’s starting
value. Each connecting line extends from the right corner of one bar’s ending
value to the left corner of the adjacent bar’s starting value.
END
draws the trend lines diagonally and connects each to the adjacent bar’s ending
value. Each connecting line extends from the right corner of one bar’s ending
value to the left corner of the adjacent bar’s ending value.

Default START

Restriction The last connect line is always drawn horizontally, extending from the
right corner of the last data bar’s ending value to the left corner of the
“final” bar’s starting value.

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the trend lines that connect the bars.

Default The GraphConnectLine style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.
WATERFALLCHART Statement 863

“Line Options” on page 1349 for available line-options.

CONNECTDECREASINGATTRS=style-element | style-element (line-options) |

(line-options)
specifies the appearance of trend lines that denote a decreasing value between bars.

Default The appearance specified by the CONNECTATTRS= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

CONNECTINCREASINGATTRS=style-element | style-element (line-options) | (line-

options)
specifies the appearance of trend lines that denote an increasing value between bars.

Default The appearance specified by the CONNECTATTRS= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows bars
with each of the skins applied.

Default The DATASKIN= option value that is specified in the

BEGINGRAPH statement. If not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with the first maintenance release of SAS 9.4, the maximum
number of skinned graphical elements is limited to 200 per plot in an
overlay or prototype layout. When this limit is exceeded for a plot,
the specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.
864 Chapter 6 • Plot Statements

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, bar labels, and
trend lines, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the filled
bar area. The INITIALBARATTRS= and FINALBARATTRS= options can
be used to specify transparency for the initial and final bars. You can
combine this option with FILLATTRS=, INITIALBARATTRS=, and
FINALBARATTRS= to set one transparency for the bar outlines and trend
lines but a different transparency for the bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=STANDARD | ALL | (display-options)

specifies which bar features to display.
STANDARD
displays outlined, filled bars; connect lines; and the “final” bar.
ALL
same as STANDARD
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

FILL displays filled bars

FINALBAR displays the “final” bar
OUTLINE displays outlined bars
CONNECT Displays line segments (trend lines) connecting adjacent bar.
The connection point is determined by the CONNECT=
option.

Default STANDARD

Tips To control the appearance of the bars, use the COLORMODEL= ,

FILLATTRS= , and OUTLINEATTRS= options.

To control the appearance of the trend lines, use the CONNECTATTRS= ,

CONNECTDECREASINGATTRS= , and
CONNECTINCREASINGATTRS= options.
WATERFALLCHART Statement 865

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled transaction bars.

Defaults If the COLORGROUP= option is not specified, then the

GraphDataDefault:Color style reference.

If the COLORGROUP= option is specified, then the

GraphData1:Color–GraphDataN:Color style references.

Interaction This option’s color specification is ignored if either the

COLORMODEL= or COLORRESPONSE= option is specified. The
transparency specification is honored in that case.

Tip The DATATRANSPARENCY= option sets the transparency for the bar
fill, bar outline, and trend lines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the bar outlines
and trend lines but a different transparency for the bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

FILLTYPE=SOLID | GRADIENT
determines whether a solid or gradient fill is used in the bars.

Default SOLID

Tip The colors that are used depend on whether the COLORGROUP= option
or the COLORRESPONSE= option is also specified.

FINALBARATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the “final” bar, if displayed.

Default the GraphFinal style element

Interaction This option is ignored if the DISPLAY= option does not display the
“final” bar.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

FINALBARTICKVALUE="string"
specifies a tick value to use on the category axis when the “final” bar is displayed

Default "Final"

Interaction This option is ignored if the DISPLAY= option does not display the
“final” bar.

INITIALBARATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the “initial” bar, if displayed.

Default the GraphInitial style element

866 Chapter 6 • Plot Statements

Interaction For this option to take effect, the INITIALBARVALUE= option must
also be specified.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

INITIALBARTICKVALUE="string"
specifies a tick value to use on the category axis when the “initial” bar is displayed.

Default "Initial"

Interaction For this option to take effect, the INITIALBARVALUE= option must
also be specified.

INITIALBARVALUE=number
specifies a value for the initial bar. The initial bar’s value is used as the starting
response value for the first transaction bar.

Interactions If this option is not specified, then the initial bar is not included in the
chart and the first transaction bar is drawn from response value 0. This
is true even if an intercept value is set by the
BASELINEINTERCEPT= option.

If necessary, the response axis data range is extended to include the

initial bar value. When a logarithmic response axis is requested and
INITIALBARVALUE= specifies 0 or a negative value, the response
axis reverts to a linear axis. To restore the log axis in that case, set
INITIALBARVALUE= to a positive value.

Note The first transaction bar starts at response value 0.

WATERFALLCHART Statement 867

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The response-variable label. If a label is not defined, then the response-
variable name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the COLORGROUP= option is in effect, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template statements. The
specified name is used primarily in legend statements to coordinate the use of colors
and line patterns between the plot and the legend.

Restriction The string is case sensitive, cannot contain spaces, and must define a
unique name within the template.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar outlines.

Default The ContrastColor and LineThickness attributes of the GraphOutlines

style element.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Line Options” on page 1349 for available line-options.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data tips.
This option provides a way to add to the data columns that appear in data tips that are
specified by the TIP= option.
(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles CATEGORY, RESPONSE, COLORGROUP, and
COLORRESPONSE.
868 Chapter 6 • Plot Statements

STAT=SUM | MEAN
specifies the statistic to be computed for the RESPONSE axis.

Default SUM

Tip If you use this option with COLORRESPONSE= in SAS programs that
were written before the third maintenance release of SAS 9.4, the chart
colors and color statistic might change from the previous SAS releases. To
restore the original colors and color statistic in that case, set
COLORSTAT= in the WATERFALLCHART statement to the same statistic
that is specified in STAT=.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar. If this
option is used, then it replaces all of the information that is displayed by default.
Roles for columns that do not contribute to the waterfall chart can be specified along
with roles that do.
(role-list)
an ordered, space-separated list of unique WATERFALLCHART and user-
defined roles. WATERFALLCHART roles include CATEGORY , RESPONSE ,
COLORGROUP , and COLORRESPONSE .

Example The following example displays data tips only for the column that is
assigned to the RESPONSE role:

TIP=(RESPONSE)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in the
data tip information: CATEGORY and RESPONSE .

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that has the IMAGEMAP option
specified, and you must write the output to the ODS HTML
destination.

Interaction This option is ignored when the plot statement is in an OVERLAY or

PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control the
formats of columns that appear in data tips.
(role-format-list)
a space-separated list of role-name = format pairs.

Example TIP=(RESPONSE)
TIPFORMAT=(RESPONSE=DOLLAR12.)
WATERFALLCHART Statement 869

Default The column format of the column assigned to the role or BEST6 if no
format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control the
labels of columns that appear in data tips.
role-label-list
a space-separated list of rolename ="string" pairs.

Example TIP=(RESPONSE)
TIPLABEL=(RESPONSE="Average Sales")

Default The column label or column name of the column assigned to the role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

URL=string-column
specifies an HTML page to display when a bar is selected.
string-column
specifies a column that contains a valid HTML page reference (HREF) for each
bar that is to have an active link.

Example http://www.sas.com/technologies/analytics/
index.html

Requirements To generate selectable bars, you must include an ODS GRAPHICS

ON statement that has the IMAGEMAP option specified, and you
must write the output to the ODS HTML destination.

For non-grouped data, the values of the column are expected to be

same for each unique RESPONSE value. If they are not, then the
results might be unpredictable.

For grouped data, the values of the column are expected to be the
same for each unique RESPONSE and group combination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some RESPONSE values, meaning
that no action is taken when the bars for those RESPONSE values
are selected.
870 Chapter 6 • Plot Statements

The URL value can be the same for different RESPONSE values,
meaning that the same action is taken when the bars for those
RESPONSE values are selected.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the secondary
X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the secondary Y2
(right) axis.

Default Y

Interactions This option is ignored if the RESPONSE= argument is not specified.

The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 875.

Details
The input data for the WATERFALLCHART statement is raw, unsummarized input data,
and the statement calculates appropriate summarization statistics (sum or mean). By
default, the bars in the chart appear in the order in which the CATEGORY values are
present in the input data. A waterfall chart is typically used to show credit and debit
transactions or successive changes to a given state.
In a waterfall chart, the bars that are calculated from the data are called “transaction”
bars. The transaction bars represent the values of the RESPONSE variable across a
series of intermediate values for the specified CATEGORY variable. You can manage
the color of the transaction bars using the COLORGROUP , COLORMODEL , or
COLORRESPONSE= option.
A waterfall chart can also display an “initial” bar and a “final” bar. The value of the
initial bar determines the starting response value for the first transaction bar. To set the
initial value, use the INITIALBARVALUE= option. If the initial bar is not displayed,
then the first transaction bar has a starting response value of 0. The value of the final bar
is set automatically to the ending value of the last transaction bar.

Example: WATERFALLCHART Statement

The following graph was generated by the “Example Program” on page 871:
Example: WATERFALLCHART Statement 871

Example Program
data transactions;
input ID $ Amount type $;
datalines;
Alpha 2000 credit
Beta -2500 debit
Gamma -2000 debit
Delta -500 debit
Epsilon 2250 credit
;
proc template;
define statgraph waterfallchart;
begingraph;
layout overlay;
waterfallchart category=id response=amount /
colorgroup=type
initialbarvalue=1000
name="waterfall";
discretelegend "waterfall";
endlayout;
endgraph;
end;
run;

proc sgrender data=transactions template=waterfallchart;

run;
872 Chapter 6 • Plot Statements
873

Part 5

Plot Axes

Chapter 7
Axis Features in Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875

Chapter 8
Axis Options in Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 889
874
875

Chapter 7
Axis Features in Layouts

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875
How Axis Features Are Determined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875
Plot Data Are Mapped to a Designated Axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 876
When Plots Share Data and a Common Axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 880
Plot Axis Types Must Agree on Common Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . 883
Controlling Axis Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883
Setting the Axis Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 884
Adjusting the Axis View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 884
Adjusting Axis Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 885
Adjusting Axis Offsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 886

Overview
GTL plots are specified within layout blocks that enable you to control the graph display
features, including the display of the axes for the plots within the layout. For example,
the LAYOUT OVERLAY statement has XAXISOPTS= and YAXISOPTS= options that
enable you to specify axis features for the plots within the layout.
The following sections explain how the axis features are determined in a layout. The
sections also discuss the issues that you must consider when setting the axis type or
adjusting the appearance of the axis display.

How Axis Features Are Determined

Overview
The GTL uses various criteria to determine the displayed axis features for a graph.
Generally, axis features are based on the following criteria:
• the layout type
• the order of plot statements in the layout and the options specified on those
statements
876 Chapter 7 • Axis Features in Layouts

• the use of primary and secondary axes on the plots (when secondary axes are
supported)
• the plot type
• the column or columns of data that contribute to defining the axis range
• the data formats for the contributing data columns
Because the default axis features depend on a combination of the factors above, it is
useful to understand how the axis features are determined in the templates that you
build:
• how the data are mapped to the plot axes
• how the various layout types manage the axes.

Plot Data Are Mapped to a Designated Axis

Overview for Axis Mapping

Depending on the layout type and the plots that you specify within the layout, you can
manage up to four axes for two-dimensional plots:
• a primary X (bottom) axis
• a primary Y (left) axis
• a secondary X axis (X2 or “top” axis)
• a secondary Y axis (Y2 or “right” axis)
Within single-cell layouts (for example, OVERLAY layout), there can be just one each
of these four axes. However, within multi-cell layouts (for example, LATTICE layouts),
each cell can display the four axes. Thus, there can be multiple X, X2, Y, and Y2 axes
across the columns and rows in the layout grid. In a lattice-type layout, you might have
to use layout options to specify how the data ranges and axis display should be managed.
This section discusses the simpler case for axis mapping, and “Axis Mapping in Lattice-
type Layouts” on page 877 discusses the case for lattice-type layouts.
Note: GRIDDED layouts can be used to create a grid of cells, but the cells are
independent. Thus, axes in the grid cannot be managed across columns and rows, so
the simpler case applies to GRIDDED layouts.

Primary and Secondary Axes

By default, plot data are mapped to the primary axes. To enable you to override the
default, plot statements that support a secondary X2 axis provide an XAXIS= option that
can map data to the X or X2 axis. Plot statements that support a Y axis provide a
YAXIS= option that can map data to the Y or Y2 axis.
To determine the axis features within a layout, the GTL must first determine what data
must be mapped to a particular axis. Thus, your use of primary and secondary axes on
plot specifications affects the GTL’s determination of default axis features for the layout.
For example, the plot statements in the following template specify Y-data mappings to
the Y2 and Y axes:
proc template;
define statgraph y2axis;
begingraph;
layout overlay;
histogram weight / scale=count yaxis=y2 ;
How Axis Features Are Determined 877

histogram weight / scale=percent yaxis=y ;

densityplot weight / normal();
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.class template=y2axis;
run;

In this example, the first HISTOGRAM maps its Y-axis data to the Y2 axis, and the
second HISTOGRAM maps its Y-axis data to the Y axis. The DENSITY plot does not
explicitly map its Y-axis data, so the default Y axis is used. None of the plots explicitly
maps X-data, so the default X axis is used for all three plots. Thus, the GTL must
manage any interactions that result from representing multiple plots on the X and Y
axes. For example, on the X axis, it must determine an appropriate data range for
representing the data values of all three plots.
When establishing axis features for each axis, the GTL determines which plot
specifications map data to the axis. The GTL also collects the data for all of the plots
that must be represented and maps that data to the designated axis. “When Plots Share
Data and a Common Axis” on page 880 discusses the criteria the GTL uses to
determine the axis features for the axes after this mapping has been done for each axis.

Axis Mapping in Lattice-type Layouts

Lattice-type layouts (LAYOUT LATTICE, LAYOUT DATALATTICE, and LAYOUT
DATAPANEL) present a grid of graphs that automatically aligns plot areas and tick
display areas across grid cells. This alignment facilitates data comparisons among
graphs, and for those comparisons to be meaningful, the graph axes must be coordinated
across the columns and rows in the grid. All of the principles discussed in “Overview for
Axis Mapping” on page 876 apply to the lattice-type layouts. In addition, because there
can be multiple X, X2, Y, and Y2 axes across grid cells, you might have to use layout
options to specify how the data ranges and axis display should be managed.
878 Chapter 7 • Axis Features in Layouts

For example, the following template uses a LAYOUT LATTICE to generate a grid that
displays a height analysis next to a weight analysis. By default in a LAYOUT LATTICE
statement, the options ROWDATARANGE= and ROW2DATARANGE= are set to
DATA. The DATA setting scales the Y-axis and Y2-axis data ranges separately for each
cell in the layout. To ensure that the Y-axis data range is the same in both cells, the
example specifies ROWDATARANGE=UNION. Similarly, to ensure that the Y2-axis
data range is the same in both cells, the example specifies
ROW2DATARANGE=UNION:
proc template;
define statgraph y2axis;
begingraph;
layout lattice / columns=2 columngutter=10
rowdatarange=union row2datarange=union ;
layout overlay;
histogram height / scale=count yaxis=y2;
histogram height / scale=percent yaxis=y;
densityplot height / normal();
endlayout;
layout overlay;
histogram weight / scale=count yaxis=y2;
histogram weight / scale=percent yaxis=y;
densityplot weight / normal();
endlayout;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=y2axis;

run;

By default in a LAYOUT LATTICE statement, the options COLUMNDATARANGE=

and COLUMN2DATARANGE= are also set to DATA. But in this analysis, the height is
How Axis Features Are Determined 879

a separate measure from the weight, so the separate scales are appropriate for the X-axes
across cells. If the X-axes were displaying the same measure (for example, comparing
the height of female subjects to the height of male subjects), then you could specify
COLUMNDATARANGE=UNIONALL. This would set the same scaling to the X-axis
data ranges across the two layout columns. In this example you would not bother
changing the default COLUMN2DATARANGE= setting because the X2 axis is not
needed.
Note: For DATALATTICE and DATAPANEL statements, UNIONALL is the default
value for the data ranges. Thus, you would not have to change the data ranges unless
you wanted to set UNION to scale data ranges per row or per column in the layout.
In the example, scaling the data ranges across the row ensures proper axis scaling.
However, the graph display is cluttered by the duplicate display of ticks, axis values, and
axis labels on both the Y and Y2 axes. To simplify the display, you can consolidate the
axes. To do so, use a ROWAXES block to display a single Y axis for both cells, and a
ROW2AXES block to display a single Y2 axis for both cells. The consolidated view
removes the internal axes from the grid and displays only the external axes:
proc template;
define statgraph y2axis;
begingraph;
layout lattice / columns=2 columngutter=10
rowdatarange=union row2datarange=union;
rowaxes;
rowaxis / griddisplay=on;
endrowaxes;
row2axes;
rowaxis;
endrow2axes;
layout overlay;
histogram height / scale=count yaxis=y2;
histogram height / scale=percent yaxis=y;
densityplot height / normal();
endlayout;
layout overlay;
histogram weight / scale=count yaxis=y2;
histogram weight / scale=percent yaxis=y;
densityplot weight / normal();
endlayout;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=y2axis;

run;
880 Chapter 7 • Axis Features in Layouts

When using ROWAXES or ROW2AXES blocks in a LATTICE layout, you nest within
the block one ROWAXIS statement for each row in the layout grid. The ROWAXIS
statements are applied sequentially to the rows, and each ROWAXIS statement specifies
the axis options for the Y or Y2 axes in its corresponding row. ROWAXIS statements
within the ROWAXES block apply to the Y axes, and ROWAXIS statements within the
ROW2AXES block apply to the Y2 axes. This example has just a single row in the grid,
so each block specifies only one ROWAXIS statement. Notice that the ROWAXIS
statement in the ROW2AXES block does not use any options. Thus, it consolidates Y2
axes in the row into a single, external Y2 axis, but it does not alter the default features of
that axis. For columns in the grid, the LATTICE layout provides COLUMNAXES and
COLUMN2AXES blocks. These blocks use COLUMNAXIS statements to externalize X
and X2 axes and specify their features.
When you use DATALATTICE and DATAPANEL layouts, the layout dynamically
generates a grid that contains as many cells as can be produced from the combination of
classification values. In those layouts the axes are always external, and you can use the
COLUMNAXISOPTS=, COLUMN2AXISOPTS=, ROWAXISOPTS=, and
ROW2AXISOPTS= options to specify the features for the axes. The settings on each
option apply across the entire grid. For example, if you specify the ROWAXISOPTS=
option in a DATALATTICE layout, then the specified settings apply to the external Y
axes in every row.

When Plots Share Data and a Common Axis

Overview
If a layout block contains multiple plots that share data and a common axis, then the plot
settings often interact in ways that affect the axis features. Axis features include the axis
type, axis label, tick-mark layout, and so on. The GTL resolves these interactions in
ways that vary according to the layout block and plot statements.
Note: Axis interactions might not occur if other settings in the template prevent them.
As discussed in “Plot Data Are Mapped to a Designated Axis” on page 876, if two
How Axis Features Are Determined 881

plot statements are within an OVERLAY layout, then one of them might map its data
to the X axis and the other might map its data to the X2 (top) axis. Mapping to
separate axes can avoid the interactions that might occur if they both mapped their
data to the X axis.

Axis Features in Overlay-type Layouts

Overlay-type layouts (OVERLAY, OVERLAYEQUATED, and PROTOTYPE, for
example) build a composite from one or more GTL-statements.
Within overlay-type layouts, if you do not explicitly set axis features in your template
statements, then the GTL automatically determines them. It sets the axis features based
on the layouts and plots in the layout block and the data that are associated with the
template at run time.
If only one plot statement within an overlay-type layout generates an axis, then
determining axis features is straight forward: the features are derived directly from the
plot type and the columns that are used for the plot data. For example, if a LAYOUT
OVERLAY block contains a single SCATTERPLOT and the X= argument specifies a
numeric column of children’s weights, then the default X-axis type is LINEAR. The
default X-axis label is the column label of the Weight column. If the Weight column has
no defined label, then the column name is used as a label.
When an overlay-type layout contains multiple plots that generate axes, the GTL can
determine default axis features for the shared axes. Alternatively, you can use the
PRIMARY= option on one of the plot statements to specify which plot you want the
GTL to use. The following code fragment explicitly specifies that the SCATTERPLOT
of children’s weights be used to determine axis features within the layout:
layout overlay;
scatterplot x=weight ... / primary=true;
...

• If no plot in an overlay-type layout is designated as primary, then the first plot that
generates an axis is considered primary on a per-axis basis.
• If PRIMARY=TRUE for a plot within an overlay-type layout, then that plot’s data
columns, data type, and plot type determine the default axis features. An explicitly
specified primary plot determines the default axis features regardless of where that
plot statement occurs within the layout block.
• Only one plot can be primary on a per-axis basis. If multiple plots specify
PRIMARY=TRUE for the same axis, then the last one encountered is considered
primary.
The following SCATTERPLOT specifies a character column on the X= argument:
layout overlay;
scatterplot x=name ... / primary=true;
...

In this case, the default X-axis type is DISCRETE and the X-axis label is the label that is
assigned to column Name, or Name if no label is assigned to column Name.
Note: The SAS format on the primary plot’s column determines the axis format,
although the axis might not use that SAS format “as-is” from the column.
If a SCATTERPLOT’s X= argument specifies a column that has a SAS DATETIME
format, then the default X-axis type is TIME. The default X-axis label is the column
label or name of the DateTime column:
layout overlay;
scatterplot x=date ... / primary=true;
882 Chapter 7 • Axis Features in Layouts

...

For some plot types, the default axis type does not directly correlate to the specified
column’s data type. For example, the following code fragment specifies a BARCHART
for the numeric column Age:
layout overlay;
barchart category=age ... / primary=true;
...

Because a BARCHART requires a discrete X axis, the default X-axis type in this case is
DISCRETE, in spite of the fact that column Age is numeric. The X-axis label is the
column label of Age, or the column name if no label exists.
Finally, consider a HISTOGRAM that is set as the primary plot in the layout and that
bins data values:
layout overlay;
histogram weight / binaxis=true primary=true;
...

In this case, the default X-axis type is LINEAR, but the histogram’s data bins are used
by default as the basis for the axis tick marks.

Axis Features in Data Panel and Data Lattice Layouts

The criteria discussed in “Axis Features in Overlay-type Layouts” on page 881 apply to
determining the default axis features for the plots within DATAPANEL and
DATALATTICE layouts. Both of these layout types nest a LAYOUT PROTOTYPE
statement within their layout blocks. In both cases, the plot statements within the
LAYOUT PROTOTYPE block—an overlay-type layout—determine the axis features for
the plot display.

Axis Features in Lattice-type Layouts

The LAYOUT LATTICE statement can create a grid of graphs that automatically aligns
plot areas, data display areas, labels, and headers across the columns and rows in the
layout. The layout gives you the option of unifying the scale of the data ranges that are
displayed in the graphs.
If a LAYOUT LATTICE specification generates only one cell, then no competition
exists between cells for determining axis features in the display. In this case, the axis
features are derived directly from the plot type and the columns used for the plot data.
Similarly, for multi-cell displays, if any or all of the options COLUMNDATARANGE=,
COLUMN2DATARANGE=, ROWDATARANGE=, or ROW2DATARANGE= use the
DATA setting to scale axis data ranges separately for each cell in the layout, then the
layout cells are data-independent. The data-independent cells do not interact with each
other for determining the axis features in the display.
Axes are shared in the layout when one of the options COLUMNDATARANGE=,
COLUMN2DATARANGE=, ROWDATARANGE=, or ROW2DATARANGE= is used
to unite axis data ranges for layout cells. By default in those cases, the first cell that is
drawn (by default, the top left cell) determines the axis features in the display. When
UNIONALL is in effect, those same features are used in all of the grid’s layout cells.
When UNION is in effect, those same features are used on a per-row or per-column
basis. If you specify external axes for the columns or rows in the layout, you can specify
desired axis features on the appropriate COLUMNAXIS or ROWAXIS statements used
in the layout.
For an example LATTICE layout with external axes, see “Axis Mapping in Lattice-type
Layouts” on page 877.
Controlling Axis Features 883

Axis Features in Gridded Layouts

In a GRIDDED layout the layout cells are independent of one another. Plot statements
within the layout cells do not share data and are not represented on a common axis.
Thus, no competition exists among layout cells for determining the axis features.

Plot Axis Types Must Agree on Common Axes

The GTL is extremely flexible and enables you to generate a wide variety of plot
displays. However, if you request incompatible plot displays within the same layout,
then the results are unpredictable.
“When Plots Share Data and a Common Axis” on page 880 discusses the criteria GTL
uses to determine the default axis features. After the axis type has been determined, the
GTL expects that all plots that share that axis will have the assigned axis type. The
expectation applies whether you specify axis features in your template or let GTL
determine default features.
For example, a BOXPLOT cannot be overlaid by a LINEPARM: the two types of plot
cannot share axes because the plot types are incompatible within the same set of axes.
Thus, if you were to use both a BOXPLOT statement and a LINEPARM statement
within a LAYOUT OVERLAY block, then only one of them can be displayed. The GTL
therefore displays the primary plot (the first specified plot by default, or the plot
designated as primary by setting PRIMARY=TRUE). The other plot is not displayed.
Similarly, a BARCHART requires a discrete X-axis, whereas a HISTOGRAM cannot be
displayed on a discrete axis. If you specify both a BARCHART and a HISTOGRAM
within the same overlay-type layout, then only the primary plot is displayed and the
other plot is rejected from the display.
Axis types must also be the same for plots that must share an axis across the columns or
rows in a multi-cell layout. For example, in a LAYOUT LATTICE, the GTL expects that
plots have the same axis type and data ranges if they are to share an external axis.
Otherwise, the external axis cannot be displayed for that row or column.

Controlling Axis Features

Overview
To enable you to control axis features within each of the layout types, there are different
sets of axis options for the different types of axes:

Option Category Statement

Two-dimensional axis options LAYOUT OVERLAY statement

Three-dimensional axis options LAYOUT OVERLAY3D statement

Equated axis options LAYOUT OVERLAYEQUATED statement

Lattice axis options LAYOUT LATTICE statement

DataLattice and DataPanel axes LAYOUT DATALATTICE and LAYOUT

DATAPANEL statements
884 Chapter 7 • Axis Features in Layouts

The options that are available for each layout are documented separately, but it is worth
discussing the following tasks in general for all of the layout types:
• Setting the Axis Type
• Adjusting the Axis View
• Adjusting Axis Thresholds
• Adjusting Axis Offsets

Setting the Axis Type

Within any given layout in the graph display, each plot axis is always of a particular
type. In the default cases, the axis type is always LINEAR, DISCRETE, or TIME.
The axis options for each layout statement include a TYPE= option that enables you to
specify an axis type that overrides the default selection mechanisms. When you override
the default axis type, you must be sure to specify the correct axis type for the plots that
you are defining. For every plot in the template language, the documentation indicates
what axis types it supports. Plots statements that are specified in the template are ignored
if they are incompatible with the axis type.
Each axis type has features specific to that type, and the following axis options enable
you to specify features for the different types:
LINEAROPTS=(linear-suboptions)
DISCRETEOPTS=(discrete-suboptions)
TIMEOPTS=(time-suboptions)
LOGOPTS=(log-suboptions)

One or more of these options can be specified for an axis, but the specified settings are
applied only to the axis type that supports them.
For example, a bar chart has two axes – a TYPE=DISCRETE axis for the X axis and a
TYPE=LINEAR axis for the Y axis. If a numeric column (for example, Age) is assigned
to the X role, then this column’s values are always treated as discrete values, never as a
continuous range of values. You cannot request another axis type for the X axis, but you
can request a different axis type for the Y axis.
Sometimes you want a specialized axis type depending on the nature of the data. For
example, if the data have a very large range of values (orders of magnitude apart), then
you could request that the values be displayed on a logarithmic scale. To set a
logarithmic scale, use the TYPE=LOG axis option.
Time series data benefit from displaying the X axis with a TYPE=TIME axis. A TIME
axis type requires that the column values are SAS Date, Time, or Datetime values.
Three-dimensional plots such as BIHISTOGRAM3DPARM and SURFACEPLOTPARM
always use TYPE=LINEAR for X, Y, and Z axes.
Note: Certain plot types or layouts might impose restrictions on what type of axis can be
assigned. The documentation for each plot and layout type identifies any restrictions
that might apply to the axes.

Adjusting the Axis View

The VIEWMIN= and VIEWMAX= axis options can be used to adjust the view of an
axis. You can specify minimum data values to include in the display, maximum data
values, or both (the specified values might be adjusted by the threshold calculation). By
Controlling Axis Features 885

default, the VIEWMIN= value is the minimum data value for the specified axis and the
VIEWMAX= value is the maximum data value for the specified axis.
A VIEWMIN= value that is greater than the data minimum or a VIEWMAX= value that
is less than the data maximum acts like a “zoom in” operation. The adjusted view
reduces the range of values represented on the axis and can sometimes exclude markers,
lines, or fills that would normally appear.
A VIEWMIN= value that is less than the data minimum or a VIEWMAX= value that is
greater than the data maximum acts like a “zoom out” operation. The adjusted view
extends the range of values represented on the axis and sometimes compresses the
markers, lines, or fills into a smaller area.
The following figure shows how the view settings can affect the tick and data displays.

Adjusting Axis Thresholds

On a continuous, linear axis, the THRESHOLDMIN= and THRESHOLDMAX= axis
options can be used to set a bias for including one more tick mark outside of either end
of the data range (or VIEWMIN to VIEWMAX range). The threshold range is from 0
(do not include the tick mark) to 1 (include the tick mark). The default is 0.30. The bias
at the minimum end of the axis is calculated using the THRESHOLDMIN= value and
the minimum data value (by default) or the VIEWMIN= value (if set).
The bias at the maximum end of the axis is calculated using the THRESHOLDMAX=
value and the maximum data value (by default) or the VIEWMAX= value (if set).
Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents the tick marks
from extending beyond the data range. Specifying THRESHOLDMIN=1 and
THRESHOLDMAX=1 ensures that the data range is bounded by tick marks.
The following figure shows how the threshold settings can affect the tick display on an
axis. In the figure, 8 is the minimum value for the display and 29 is the maximum value.
886 Chapter 7 • Axis Features in Layouts

Adjusting Axis Offsets

The OFFSETMIN= and OFFSETMAX= axis options can be used to reserve an area at
the minimum end of an axis, the maximum end, or both ends. No tick marks are
displayed in the reserved areas.
The offset range is from 0 to 1, and the specified value is used to calculate the offset as a
percentage of the full axis length. The larger the offset area that is reserved, the less
space is available for the tick display area. The default offset reserves just enough area to
fully display markers and other graphical features near the ends of an axis.
The following figure shows how offset values of 0.08 might compare with the default
offsets for a continuous axis.

This next figure shows how offset values might affect the discrete axis of a bar chart.
Controlling Axis Features 887
888 Chapter 7 • Axis Features in Layouts
889

Chapter 8
Axis Options in Layouts

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 889
Axis Options for LAYOUT OVERLAY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 889
Axis Options for LAYOUT OVERLAY3D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 945
Axis Options for LAYOUT LATTICE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 963
Axis Options for LAYOUT OVERLAYEQUATED . . . . . . . . . . . . . . . . . . . . . . 1013
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL . . . . 1032

Dictionary

Axis Options for LAYOUT OVERLAY

Axis options for the plots within an OVERLAY layout.
Interaction: The OVERLAY’s axis options are ignored when the LAYOUT OVERLAY statement is
nested within another layout type that has external axes in effect. For example, the
axis options are ignored when the OVERLAY is nested in a LAYOUT LATTICE with a
COLUMNAXIS= or ROWAXIS= option in effect.
Note: Unless otherwise indicated in an option description, each axis option is available for
the X, Y, X2, and Y2 axis.
See: “LAYOUT OVERLAY Statement” on page 136

Syntax
Axis options for the plots within an OVERLAY layout are specified with the following
options on a LAYOUT OVERLAY statement:
XAXISOPTS=(axis-options)
YAXISOPTS=(axis-options)
X2AXISOPTS=(axis-options)
Y2AXISOPTS=(axis-options)

General Options for All Axes in an Overlay

The options that are documented in this section can be used with any of the axis types
that are supported within an OVERLAY layout. Subsequent sections in the chapter
890 Chapter 8 • Axis Options in Layouts

document the axis options that are available only for specific axis types: discrete, linear,
log, or time axes. The following table provides a summary of the options.

Statement Option Description

DISCRETEOPTS Specifies options for a discrete axis.

DISPLAY Controls which axis features are displayed on

the primary axis.

DISPLAYSECONDARY Controls which axis features are displayed on

the secondary axis.

GRIDATTRS Specifies the attributes of the grid lines.

GRIDDISPLAY Specifies whether axis grid lines are

displayed.

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the

axis label.

LABELFITPOLICY Specifies a policy for fitting axis labels in the

available space.

LABELPOSITION Specifies the position of the axis label.

LABELSPLITCHAR Specifies one or more characters on which the

axis labels can be split, if needed.

LABELSPLITCHARDROP Specifies whether the split characters should

be included in the displayed axis labels.

LABELSPLITJUSTIFY Specifies the justification of the strings that

are inside the axis label blocks.

LINEAROPTS Specifies features for a standard numeric

interval axis.

LINEEXTENT Specifies the extent of the axis line.

LOGOPTS Specifies features for a log axis.

NAME Assigns a name to an axis for reference in

other statements.

OFFSETMAX Reserves an area at the maximum end of the

axis. No tick marks are displayed in the
reserved area.

OFFSETMIN Reserves an area at the minimum end of the

axis. No tick marks are displayed in the
reserved area.
Axis Options for LAYOUT OVERLAY 891

Statement Option Description

REVERSE Specifies whether the tick values should

appear in the reverse order.

SHORTLABEL Specifies an alternate axis label to use if the

default or specified axis label is too long for
the axis length.

TICKSTYLE Specifies the placement of tick marks in

relation to the axis line.

TICKVALUEATTRS Specifies the color and font attributes of the

axis tick values.

TICKVALUEHALIGN Specifies the horizontal alignment for all of

the tick values that are displayed on the Y and
Y2 axes.

TICKVALUEVALIGN Specifies the vertical alignment for all of the

tick values that are displayed on the X and X2
axes.

TIMEOPTS Specifies features for a TIME axis.

TYPE Specifies the type of axis to use.

DISCRETEOPTS=(discrete-axis-options)
specifies one or more options for a discrete axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is space
separated.

Interaction This option is ignored if the axis type is not DISCRETE.

See “Options for Discrete Axes Only” on page 903 for the options that
you can use for discrete-axis-options.

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on the primary axis.
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed
NONE
specifies that no axis features are displayed
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL displays the axis label

LINE displays the axis line
TICKS displays the tick marks
892 Chapter 8 • Axis Options in Layouts

TICKVALUES displays the values that are represented by the major tick
marks

Default STANDARD

Tips The default line attributes for the axis line and axis tick marks are defined
in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall chart
can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

DISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which axis features are displayed on the secondary axis. A secondary axis is
not an independent axis. Rather, it mirrors the primary axis. Thus, for this option to
take effect, all plot statements in the layout must map data to the same primary axis.
For example, a secondary X2 axis can be displayed on top in the layout, provided all
plot statements set XAXIS=X to map data to the primary X axis (bottom). Similarly,
a secondary Y2 axis can be displayed to the right in the layout, provided all plot
statements set YAXIS=Y to map data to the primary Y axis (left).
NONE
specifies that no axis features are displayed
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed on the
secondary axis
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL displays the axis label

LINE displays the axis line
TICKS displays the tick marks
TICKVALUES displays the values that are represented by the major tick
marks

Default NONE

Restriction If some plot statements set XAXIS=X and others set XAXIS=X2, both
the X and X2 axis are primary and a secondary X axis cannot be
displayed. In that case, this option is ignored. The same applies for the
Y axes.

Tip Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis
grid lines.
Axis Options for LAYOUT OVERLAY 893

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tip On a log axis, this option affects the appearance of the major grid lines
only. It does not affect the appearance of the minor grid lines. To
control the appearance of the minor grid lines on a log axis, use the
MINORGRIDATTRS= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template to
absolutely control the display of grid lines or to allow interaction with the current
style to decide whether grid lines are displayed.
AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element in
the current style contains DisplayOpts="ON."
AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in the
current style contains DisplayOpts="OFF."
ON
specifies that grid lines are always displayed. The current style has no override.
OFF
specifies that grid lines are never displayed. The current style has no override.
The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines style
element. Most supplied templates use the default setting AUTO_OFF to indicate a
preference for not displaying grid lines, but allowing the style to override.

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no

ON any value yes

894 Chapter 8 • Axis Options in Layouts

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style has
no preference about grid lines and the graphics template setting for grid
lines is always used.

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The list
form implies that all included string literals or dynamics will be concatenated.

Default The default label is derived from the primary plot in the layout. For
more information, see “When Plots Share Data and a Common Axis”
on page 880.

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the axis label is too long to fit along the axis, then it is truncated by
default.

Tip Use the SHORTLABEL= option to specify an alternate axis label to be

used whenever truncation would normally occur.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display the axis label.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LABELFITPOLICY=AUTO | SPLIT | SPLITALWAYS

specifies a policy for fitting axis labels in the available space.
AUTO
uses the short label, when specified, instead of the original label. If the short label
does not fit, then it is clipped. When no short label is specified, the original label
is clipped.
SPLIT
splits the axis label at a split character, which is specified by the
LABELSPLITCHAR= option, only when necessary in order to make the label fit
the available space. The short label is not used. A split does not occur at a split
character if a split is not needed at that location. If the label does not contain any
of the specified split characters, then it is not split. A label that cannot be split or
Axis Options for LAYOUT OVERLAY 895

that does not fit the available space even after splitting might overlap the
adjoining space.
SPLITALWAYS
always split the axis label at every occurrence of a split character, which is
specified by the LABELSPLITCHAR= option. If the label cannot be split, then it
is clipped.

Default AUTO

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

When the overlay layout is nested in a lattice layout, SPLIT is ignored

and AUTO is used instead.

Note When LABELPOSITION=CENTER, the available area is the full

axis, including the axis offsets. When
LABELPOSITION=DATACENTER, the available area is the tick
display area, excluding the axis offsets.

LABELPOSITION=CENTER | DATACENTER | TOP | BOTTOM | LEFT |

RIGHT
specifies the position of the axis label.
CENTER
centers the axis label in the axis area. For the Y and Y2 axes, the label is oriented
vertically and is centered in the axis area (including the offsets). The label is
positioned to the left of the tick values for the Y axis or to the right of the axis
values for the Y2 axis. For the X and X2 axes, the label is centered in the axis
area (including the offsets). It is positioned below the tick values for the X axis or
above the axis values for the X2 axis.
DATACENTER
centers the axis label in the axis tick display area. For the Y and Y2 axes, the
label is oriented vertically and is centered in the axis tick display area (excluding
the offsets). It is positioned to the left of the tick values for the Y axis or to the
right of the axis values for the Y2 axis. For the X and X2 axes, the label is
centered in the axis tick display area (excluding the offsets). The label is
positioned below the tick values for the X axis or above the axis values for the
X2 axis.
TOP | BOTTOM
orients the label horizontally at the top or bottom of the axis area. The label is
right-justified in the axis area for the Y axis and left-justified for the Y2 axis. If
there is not sufficient room in the axis area to display the label, then the label
grows to the right for the Y axis and to the left for the Y2 axis.

Restriction These options are valid for the Y and Y2 axes only.

Note When TOP or BOTTOM is used, the label might collide with other
graphical features.

LEFT | RIGHT
positions the label to the left or right of the axis area. The label is centered
vertically in the axis area.

Restriction These options are valid for the X and X2 axes only.
896 Chapter 8 • Axis Options in Layouts

Note When LEFT or RIGHT is used, the label might collide with other
graphical features.

The following figure shows the CENTER and DATACENTER positions for a blue Y
axis label Type and a red X axis label MPG.

In this example, an axis offset is applied to the maximum end of both axes in order to
demonstrate the difference between CENTER and DATACENTER. CENTER
centers the labels on the entire axis area, including the offset. DATACENTER
centers the labels on the tick display areas, which does not include the offset.
The next figure shows the TOP and LEFT positions, and the BOTTOM and RIGHT
positions for the same axis labels.

Default CENTER

Restriction This option does not support collision avoidance. In some cases, axis
label collisions can occur in the axis area.

Interaction When LEFT, RIGHT, TOP, or BOTTOM is in effect, the

SHORTLABEL= option is ignored.

See SHORTLABEL= on page 900 for information about how short labels
are used.

LABELSPLITCHAR="character-list"
specifies one or more characters on which the axis labels can be split, if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
axis label. In that case, all of the specified split characters together are treated as a
single split character.
Axis Options for LAYOUT OVERLAY 897

When LABELFITPOLICY=SPLIT, if the axis label does not fit the available space,
then it is split on a specified split character only if a split is needed at that point to
make the label fit. In this case, a split might not occur on every split character. When
LABELFITPOLICY=SPLITALWAYS, the axis label is split unconditionally on
every occurrence of a split character. If the axis label does not contain any of the
specified split characters, the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
labelsplitchar="abc"

The LABELSPLIT=TRUE option must also be specified.

Interactions This option has effect only when LABELPOSITION= is CENTER

or DATACENTER.

The LABELSPLITCHARDROP= option specifies whether the split

characters are included in the displayed data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the LABELSPLITJUSTIFY= option to specify the justification

of the strings in the axis label block.

LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed axis labels.
TRUE
drops the split characters from the axis label display.
FALSE
includes the split characters in the axis label display. When
LABELSPLIT=TRUE and LABELSPLITCHARDROP=FALSE, each split
character remains as the last character in the current line. The characters that
follow the split character, up to and including the next split character, are then
wrapped to the next line.

Default TRUE. The split characters are dropped from the axis label.

Requirement The LABELSPLIT=TRUE option must also be specified.

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHAR= option specifies the split characters.

898 Chapter 8 • Axis Options in Layouts

See “boolean ” on page 1339 for other Boolean values that you can use.

LABELSPLITJUSTIFY=justification
specifies the justification of the strings that are inside the axis label blocks.
justification
CENTER | LEFT | RIGHT
specifies the justification for the X or X2 axis label.
CENTER | TOP | BOTTOM
specifies the justification for the Y or Y2 axis label.

Default CENTER

Requirement LABELSPLIT=TRUE option must also be specified.

Interaction This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

LINEAROPTS=(linear-axis-options)
specifies one or more options for a numeric interval axis. Options must be enclosed
in parentheses. Each option is specified as a name = value pair and each pair is space
separated.

Interaction This option is ignored if the axis type is not LINEAR.

See “Options for Linear Axes Only” on page 912 for the options that you
can use for linear-axis-options.

LINEEXTENT=FULL | DATA | number

specifies the extent of the axis line.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
FULL
specifies an axis line that extends along the entire length of the axis.
DATA
specifies an axis line that extends through the data range from the minimum
offset to the maximum offset.
number
specifies how much the axis line extends from DATA toward FULL as a decimal
proportion. A value of 0 is equivalent to DATA, and a value of 1 is equivalent to
FULL.

Range 0–1

Tip A numeric value is useful for bar charts when DATA terminates the axis
line at the midpoint positions of the minimum and maximum bars. In
that case, you can specify a numeric value to lengthen the axis line so
that it extends to the full width of both bars.
Axis Options for LAYOUT OVERLAY 899

The following figure shows a simple example of each value for the X and Y axis lines.
The light-blue dashed lines depict the minimum and maximum offsets that are set on the
axes.

Default FULL

Restriction This option is valid only in OVERLAY and OVERLAYEQUATED

layouts.

Interaction This option overrides the AXISLINEEXTENT= option in the

BEGINGRAPH statement.

Tip The graph wall outline might appear to be an axis line. In that case, use
the WALLDISPLAY=NONE or WALLDISPLAY=(FILL) option in the
layout statement to suppress the wall outline.

LOGOPTS=(log-axis-options)
specifies one or more options for a log axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is space
separated.

Interaction This option is ignored if the axis type is not LOG.

See “Options for Log Axes Only” on page 925 for the options that you
can use for log-axis-options.

NAME="string"
assigns a name to an axis for reference in other statements. Currently, it is used only
in an AXISLEGEND statement.

Interactions This option is ignored unless the axis is discrete. The axis can be
discrete by default, or explicitly set to discrete with a TYPE=
DISCRETE setting.

For this option to take effect, an axis legend must be enabled. To

enable an axis legend, the DISCRETEOPTS= option must set the
TICKVALUEFITPOLICY to either EXTRACT or
EXTRACTALWAYS. In addition, an AXISLEGEND statement must
be specified to generate the axis legend.

OFFSETMAX=AUTO | AUTOCOMPRESS | number

reserves an area at the maximum end of the axis. No tick marks are displayed in the
reserved area.
AUTO
reserves just enough area to fully display markers and other graphical features
near the maximum end of an axis.
900 Chapter 8 • Axis Options in Layouts

AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.
number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886

OFFSETMIN=AUTO | AUTOCOMPRESS | number

reserves an area at the minimum end of the axis. No tick marks are displayed in the
reserved area.
AUTO
reserves just enough area to fully display markers and other graphical features
near the maximum end of an axis.
AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.
number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886.

REVERSE=TRUE | FALSE
specifies whether tick values should appear in the reverse order.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

SHORTLABEL="string"
specifies an alternate axis label to display when the default label or the label
specified by the LABEL= option is too long to fit the available space.
When LABELPOSITION=CENTER (default), the available space for an axis label is
the full axis, including the axis offsets. When LABELPOSITION=DATACENTER,
the available space for an axis label is the axis tick display area, which excludes the
axis offsets. If the label length exceeds the available space, then the label is anchored
at the left or bottom offset. It extends beyond the opposing offset until it reaches the
end of the axis where it is truncated. An ellipsis designates the truncation.

Interactions This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

This option has effect only when the LABELPOSITION= option is set
to CENTER or DATACENTER.
Axis Options for LAYOUT OVERLAY 901

Note If the specified label is itself too long for the axis, it is truncated in the
display.

TICKSTYLE=OUTSIDE | INSIDE | ACROSS

specifies the placement of tick marks in relation to the axis line. The figure shows
the tick display for each value.

OUTSIDE displays tick marks outside of the axis frame.

INSIDE displays tick marks inside the axis frame.
ACROSS displays tick marks across the axis line.

Default The GraphAxisLines:TickDisplay style reference.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick marks.

Notes This option has no affect on the placement of the tick values, which are
always outside the axis frame.

This option applies to both major ticks and minor ticks.

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick values.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TICKVALUEHALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment for all of the tick values that are displayed on the
Y and Y2 axes.
902 Chapter 8 • Axis Options in Layouts

Defaults RIGHT for a Y axis

LEFT for a Y2 axis

Restriction This option is valid for the Y and Y2 axes only.

TICKVALUEVALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment for all of the tick values that are displayed on the X
and X2 axes.

Defaults TOP for an X axis

BOTTOM for an X2 axis

Restriction This option is valid for the X and X2 axes only.

TIMEOPTS=(time-axis-options)
specifies one or more options for a time axis.

Requirements Columns associated with a time axis must be in SAS time, SAS
date, or SAS datetime units and have an associated SAS time, date,
or datetime format.

Options must be enclosed in parentheses. Each option is specified as

a name = value pair and each pair is space separated.

Interaction This option is ignored if the axis type is not TIME.

See “Options for Time Axes Only” on page 934 for the options that
you can use for time-axis-options.
Axis Options for LAYOUT OVERLAY 903

TYPE=AUTO | DISCRETE | LINEAR | TIME | LOG

specifies the type of axis to use.
AUTO
requests that the axis type be automatically determined, based on the overlay
contents. For more information, see “When Plots Share Data and a Common
Axis” on page 880.
DISCRETE
uses a DISCRETE axis if possible. The data for discrete axes can be character or
numeric. You can add a DISCRETEOPTS= option list to customize this axis
type.
LINEAR
uses a LINEAR axis if possible. You can add a LINEAROPTS= option list to
customize this axis type.
TIME
uses a TIME axis if possible. Data for this axis must be SAS time, SAS date, or
SAS datetime values. You can add a TIMEOPTS= option list to customize this
axis type.
LOG
uses a LOG axis if possible. You can add a LOGOPTS= option list to customize
this axis type.

Interaction If a log axis is requested and the axis data contains 0 or negative
values, the axis reverts to a linear axis. This outcome can occur for
the response axis of a bar chart, line chart, needle plot, or waterfall
chart when a baseline intercept of 0 or less is specified. It can also
occur for the response axis of a waterfall chart when an initial bar
value of 0 or less is specified. To get a log response axis in those
cases, set the baseline intercept or initial bar value to a positive
value.

Default AUTO

Interactions If this option is set to anything other than AUTO, then plots within the
layout are dropped from the display if their data types or data ranges
do not match the axis type requirements. For more information, see
“Plot Axis Types Must Agree on Common Axes” on page 883.

After the axis type is determined (whether you set a specific type or
AUTO is in effect), you can use only options that are supported by
that axis type. For example, if TYPE=TIME, then only the general
OVERLAY axis options and those available on TIMEOPTS= are
supported.

Options for Discrete Axes Only

The options that are documented in this section can be used with the DISCRETEOPTS=
axis option. The following table provides a summary of the options.

Discrete Axis Options Description

COLORBANDS Specifies the display of alternating wall-color

bands corresponding to the discrete axis bins.
904 Chapter 8 • Axis Options in Layouts

Discrete Axis Options Description

COLORBANDSATTRS Specifies the appearance of the alternating

wall-color bands.

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the
TICKVALUELIST= option.

TICKTYPE Specifies the position of the axis tick mark.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision on an axis.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the list of tick values that are

displayed on the axis.

TICKVALUEROTATION Specifies how the tick values are rotated on

the X and X2 axes.

TICKVALUESPLITCHAR Specifies whether the split characters are

included in the displayed tick values.

TICKVALUESPLITCHARDROP Specifies whether the split characters are

included in or dropped from the displayed tick
value.

TICKVALUESPLITJUSTIFY Specifies justification of the strings that are

inside the tick value block.

COLORBANDS=NONE | EVEN | ODD

specifies the display of alternating wall-color bands corresponding to the discrete
axis bins.

Default NONE

Restriction This option applies to discrete axes only.

Interaction Specifying this option for more than one axis in the layout might have
unexpected results. The order in which color bands are drawn might
not match the order in which the axis options are specified.

Note The full width of a color band is the distance between midpoints. When
no axis offsets are specified, the first band begins at one-half of the
midpoint distance, and the last band ends at one-half of the midpoint
distance. When axis offsets are specified, the first and last color bands
on the axis might extend into their adjacent offsets by as much as half
the color-band width.

Tips Borders for the color bands can be added by setting TICKTYPE=
INBETWEEN in the DISCRETEOPTS= option, and by setting
GRIDDISPLAY= ON.
Axis Options for LAYOUT OVERLAY 905

Because alternating color bands are drawn on top of the plot wall, this
option can be coordinated with the LAYOUT OVERLAY statement’s
WALLCOLOR= option.

COLORBANDSATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the alternating wall-color bands. For the alternating
colors, one set uses the WALLCOLOR= colors that are set in the LAYOUT
OVERLAY statement, and the other set uses the colors set on this option.

Default The GraphBlock style element.

Restriction This option applies to discrete axes only.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string values
that are displayed on the axis in place of the values in the TICKVALUELIST=
option. The strings map one-to-one positionally with the values that are listed in the
TICKVALUELIST= option.

Default Determined by the system or by the TICKVALUELIST= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Tip This option should be used with the TICKVALUELIST= option.

The number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

Example The following example specifies the axis tick values 10, 20, 30, and
40, and the tick display values A, B, C, and D:
tickvaluelist=("10" "20" "30" "40");
tickdisplaylist=("A" "B" "C" "D");

TICKTYPE=MIDPOINT | INBETWEEN
specifies the position of the axis tick marks.

MIDPOINT places the tick marks at the midpoint value location.

INBETWEEN places the tick marks half way between adjacent midpoint
locations.

Default MIDPOINT

Restriction This option applies to discrete axes only.

Note Starting with the second maintenance release of SAS 9.4, when
TICKTYPE=INBETWEEN, the outermost tick marks and grid lines at
each end of the axis are not drawn.
906 Chapter 8 • Axis Options in Layouts

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness of a
collision-avoidance policy depends on the number of tick values, their length, and
the length of the axis. Which policies are valid depends on the axis on which this
option is used. For the Y and Y2 axes, the following policies are valid:
EXTRACT
displays consecutive integers along the axis instead of the actual tick values in
order to represent those tick values. In most cases, this policy is implemented if
the system estimates that a collision might occur. If no collision occurs, then the
actual tick values are displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User's Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User's Guide.

NONE
makes no attempt to avoid collisions between tick values. Tick values are display
even when they collide.
SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space. A split does not occur at a split character if a split is
not needed at that location. If the value does not contain any of the specified split
characters, then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR=

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that is
specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR=

SPLITALWAYSTHIN
same as SPLITALWAYS, except that thinning is performed when long words do
not fit the available space.
SPLITTHIN
same as SPLIT, except that thinning is performed when long words do not fit the
available space.
THIN
eliminates alternate tick values.
For the X and X2 axes, the following policies are valid:
Axis Options for LAYOUT OVERLAY 907

EXTRACT
display consecutive integers along the axis instead of the actual tick values to
represent those tick values. In most cases, this policy is implemented if the
system estimates that a collision might occur. If no collision occurs, then the
actual tick values are displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User's Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User's Guide.

NONE
does not attempt to fit tick values that collide.
ROTATE
rotates the tick values if a collision occurs. The TICKVALUEROTATION=
option specifies whether the values are rotated to a 45-degree diagonal or a 90-
degree vertical position. By default, the values are rotated to a 45-degree
diagonal position.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs. The
TICKVALUEROTATION= option specifies whether the values are rotated to a
45-degree diagonal or a 90-degree vertical position. By default, the values are
rotated to a 45-degree diagonal position.
ROTATEALWAYSDROP
attempts the ROTATEALWAYS policy, and then drops the tick values if
collisions still occur.
ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space. A split does not occur at a split character if a split is
not needed at that location. If the value does not contain any of the specified split
characters, then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR=

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that is
specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR=
908 Chapter 8 • Axis Options in Layouts

SPLITROTATE
attempts the SPLIT policy, and then the ROTATE policy.
STAGGER
alternates the tick values between two rows.
STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.
STAGGERTRUNCATE
attempts the STAGGER policy, and then the TRUNCATE policy.
THIN
eliminates alternate tick values.
TRUNCATE
shortens the tick values when they exceed a certain number of characters.
TRUNCATEROTATE
attempts the TRUNCATE policy, and then the ROTATE policy.
TRUNCATESTAGGER
attempts the TRUNCATE policy, and then the STAGGER policy.
TRUNCATETHIN
attempts the TRUNCATE policy, and then the THIN policy.

Defaults ROTATE for the X and X2 axes

THIN for the Y and Y2 axes

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format
specifies how to format the values for major tick marks.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Restrictions This option applies only to discrete axes.

Only character formats are supported.

Interaction This option is ignored when the axis tick values are extracted to an
axis legend. See TICKVALUEFITPOLICY=EXTRACT |
EXTRACTALWAYS on page 906.

Tip Use this option when you want to duplicate tick values on an axis.

TICKVALUELIST=(string-list)
specifies the list of tick values that are to be displayed on the axis.
string-list
a space-separated list of values, enclosed in parentheses. You must enclose each
value in the list in quotation marks.
Only the tick values that are included in the string list are displayed on the axis. The
values are displayed in the order in which they are listed. The data values that are not
in the list are dropped. The list can be a subset of the data values. It can also contain
Axis Options for LAYOUT OVERLAY 909

values that are not included in the actual data. A tick value that is not included in the
data appears on the axis, but no data is represented at its tick mark.

Requirements The list of values must be enclosed in parentheses.

Each value must be enclosed in quotation marks and separated from

adjacent values by a blank space.

Notes If the string list contains duplicate values, then the first occurrence
of the duplicated value in the list is honored and the remaining
instances are ignored.

When the values specified in the list are compared with the actual
data values, leading blanks are honored and trailing blanks are
ignored.

Tips You can use this option to subset the axis values or to display the
values in a specific order.

You can use this option to display values on the axis that are not
contained in the data.

Examples The following example specifies the axis tick values Sedan, Sports,
Wagon, and SUV:
tickvaluelist=("Sedan" "Sports" "Wagon" "SUV")

The following example specifies the axis tick values 10, 20, 30, and
40:
tickvaluelist=("10" "20" "30" "40")

TICKVALUEROTATION=DIAGONAL | VERTICAL
specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL rotates the tick values to a 45-degree diagonal position. The X

labels read left to right in a downward direction. The X2 labels
read left to right in an upward direction.
VERTICAL rotates the labels to a 90-degree vertical position. The labels are
always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for XAXISOPTS= and X2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESPLITCHAR="character-list"
specifies a list of characters on which the tick values can be split, if needed. When
multiple characters are specified, each character in the list is treated as a separate
split character unless the specified characters appear consecutively in the tick value.
In that case, all of the specified split characters together are treated as a single split
character.
When TICKVALUESPLITPOLICY=SPLIT, if a tick value collision is detected, then
the tick value is split on a split character only if necessary at that point in order to
avoid collision. In that case, a split might not occur on every split character. When
TICKVALUEFITPOLICY=SPLITALWAYS, the tick value is split unconditionally
910 Chapter 8 • Axis Options in Layouts

on every occurrence of a split character. If the tick value does not contain any of the
specified split characters, then it is not split.
"character-list"
one or more characters with no delimiter between each character.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the characters a, b, and c, use the following
option:
tickvaluesplitchar="abc"

Interactions This option is ignored unless option TICKVALUEFITPOLICY= is

set to SPLIT, SPLITALWAYS, SPLITTHIN, or
SPLITALWAYSTHIN.

The TICKVALUEFITPOLICY= option sets the policy that is used

to manage the split behavior of the tick values.

The TICKVALUESPLITCHARDROP= option specifies whether

the split characters are displayed or dropped from the display.

Notes When multiple characters are specified, the order of characters in

the list is not significant.

The split characters are case sensitive.

Tips Use the TICKVALUESPLITJUSTIFY= option to specify the

justification of the strings in the tick value block.

For the X and X2 axis tick values, use the TICKVALUEVALIGN=

option to specify the vertical alignment of the tick values.

For the Y and Y2 axis tick values, use the TICKVALUEHALIGN=

option to specify the horizontal alignment of the tick values.

Example The following example specifies a blank space, a comma, and an

underscore as split characters:
tickvaluesplitchar=" ,_"

TICKVALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters should be included in the displayed tick values.
The split characters are specified by the TICKVALUESPLITCHAR= option.
TRUE
drops the split characters from the tick value display. The following figure shows
an example in which TICKVALUESPLITCHARDROP=TRUE and three-word,
asterisk-delimited tick values are split on the asterisk character by using the
SPLITALWAYS policy.
Axis Options for LAYOUT OVERLAY 911

Notice that the asterisk delimiter is not displayed.

FALSE
includes the split characters in the tick value display. The fit policy determines
how the characters are displayed. If the display policy is SPLIT or SPLITTHIN
and TICKVALUESPLITCHARDROP=FALSE, then each tick value is split at a
split character only where a split is necessary in order to make the value fit the
available space. A split might not occur at every split character in the tick value.
At each split point, the split character remains as the last character in the current
line. The characters that follow the split character, up to and including the split
character at the next split point, are then wrapped to the following line. This
process repeats until the entire data tick value is displayed. The following figure
shows an example in which TICKVALUESPLITCHARDROP=FALSE and
three-word, asterisk-delimited tick values are split on the asterisk character by
using the SPLIT policy.

Notice that a split occurs on the first asterisk and not at the second. In this case, a
split is not needed at the second asterisk.
If the fit policy is SPLITALWAYS or SPLITALWAYSTHIN, and
TICKVALUESPLITCHARDROP=FALSE, then each tick value is split at every
instance of a split character in the value regardless of whether a split is actually
needed. Each split character remains as the last character in the current line. The
characters that follow each split character, up to and including the next split
character, are then wrapped to the next line. The following figure shows an
example in which TICKVALUESPLITCHARDROP=FALSE and three-word,
asterisk-delimited tick values are split on the asterisk character by using the
SPLITALWAYS policy.

Notice that a split occurs after each asterisk and each asterisk appears at the end
of the line. In this case, three lines are displayed.

Default TRUE

Interactions The TICKVALUESPLITCHAR= option specifies the split character

or characters.

This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUESPLITJUSTIFY=CENTER | LEFT | RIGHT

specifies justification of the strings that are inside the tick value block. The
justification is relative to an individual tick value’s display area and does not affect
the display of tick values that are not split.
912 Chapter 8 • Axis Options in Layouts

Defaults CENTER for an X or X2 axis

RIGHT for a Y axis

LEFT for a Y2 axis

Interaction This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

Options for Linear Axes Only

The options that are documented in this section can be used only with the
LINEAROPTS= axis option. The following table provides a summary of the options.

Linear Axis Option Description

INCLUDERANGES on page 913 Specifies one or more ranges for a broken

axis.

INTEGER Specifies that evenly spaced integer values are

used for tick marks.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether the minor tick marks are

displayed on the axis.

ORIGIN Specifies that the axis perpendicular to the

current axis be drawn at the indicated data
value.

THRESHOLDMAX Specifies a bias for including one more tick

mark at the maximum end of the axis.
Axis Options for LAYOUT OVERLAY 913

Linear Axis Option Description

THRESHOLDMIN Specifies a bias for including one more tick

mark at the minimum end of the axis.

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined by the
TICKVALUELIST= option.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision. Only the default policy (THIN) is
available for a Y or Y2 axis.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the order of the tick values for a

linear axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification

can extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on

the X and X2 axes.

TICKVALUESEQUENCE Specifies the tick values for a linear axis by

start, end, and increment.

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

INCLUDERANGES=(start–end <start2–end2 startN–endN …>)

specifies the ranges for a broken axis.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
start
specifies a numeric value or the keyword MIN. MIN specifies the minimum data
value.
end
specifies a numeric value or the keyword MAX. MAX specifies the maximum
data value.
914 Chapter 8 • Axis Options in Layouts

The following figure shows a linear axis, broken into ranges 50–52 and 56–73.

As shown in the figure, break lines are drawn to indicate the break in the axis.

Restrictions This option is valid for linear and time axes in an OVERLAY layout
only.

Only one axis can be broken. If this option is specified for both
axes, then it is honored for the vertical axis and ignored for the
horizontal axis.

When plots are associated with the X and X2 axes or with Y and Y2
axes, neither axis can be broken.

A binned heat map or histogram axis cannot be broken.

A broken axis is not supported in vector graphics output. When a

broken axis is specified and vector graphics output is requested, the
graph is converted into an image instead. A note indicating the
conversion is written to the SAS log. To restore the vector graphics
output in that case, remove the INCLUDERANGES= option from
the LINEAROPTS= or TIMEOPTS= option.

Requirements All of the ranges must be enclosed in parenthesis.

You must specify each range as a starting value, a hyphen, and an

ending value. You must separate adjacent ranges with a space.

Each range must be nonzero. A zero range such as 12–12 is

considered invalid.

Interactions When this option is specified, axis options THRESHOLDMIN=,

THRESHOLDMAX=, VIEWMIN=, VIEWMAX=, and
TICKVALUEPRIORITY= are ignored. Suboption
EXTRACTSCALE= of the TICKVALUEFORMAT= option is also
ignored.

When this option is specified, the plot statement TIP= and URL=
options are ignored.
Axis Options for LAYOUT OVERLAY 915

Notes When this option is specified, data-clipping might occur for the
following: plot markers and marker characters, box-plot outlier
markers, fixed-position data labels, needle plots and fringe plots in
the X direction, reference lines and drop lines on the broken axis,
axis tables, and relative bubble plots.

Curve label positions are based on the contiguous axis data range.
When curve labels are specified with a broken axis, the curve label
positions might not be ideal.

Tip Starting with the third maintenance release of SAS 9.4, you can use
the AXISBREAKTYPE= and AXISBREAKSYMBOL= options in
the BEGINGRAPH statement to display the break in the axis as
only a symbol on the axis line.

See “Creating a Broken Linear Axis” in SAS Graph Template Language:

User's Guide

Example includeranges=(10-500 1000-5000 10000-50000)

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the
TICKVALUEFORMAT= option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAM or HISTOGRAMPARM statement.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.
916 Chapter 8 • Axis Options in Layouts

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
The following figure shows the minor grid lines set to light blue, dotted lines on a linear
axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.
Axis Options for LAYOUT OVERLAY 917

Defaults Four ticks with five intervals in the first maintenance release of SAS
9.4 and earlier releases.

One tick with two intervals starting with the second maintenance
release of SAS 9.4.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

ORIGIN=number
specifies that the axis perpendicular to the current axis is drawn at the indicated data
value.
For managing origin settings, the GTL treats the X and Y axes as a pair, and the X2
and Y2 axes as a separate pair. Thus, if you set the Y-axis origin to 200, then the X
axis is drawn from that origin point. If the graph also displays an X2 axis, then it is
unaffected and does not move. Similarly, if you set an origin for the Y2 axis, then the
X2 axis moves to that origin point and the X axis is unaffected.
If you set an origin for the Y2 axis and there is no X2 axis, then the origin setting for
Y2 does not affect the graph display. That is, the X axis does not move to that origin
point.
If you set an origin for an axis and the axis has a tick value at that origin value, the
tick value is not displayed. Suppressing the tick value at the origin prevents the value
from colliding with the axis value on the perpendicular axis. However, it is possible
that the tick values on the orthogonal axes will collide.

Default The axis perpendicular to the current axis is drawn at the minimum
tick value minus the OFFSETMIN= value.

Restriction This option applies to linear axes only.

Interactions If the specified value is outside the data range for the current axis, then
the data range is extended to include the value.
918 Chapter 8 • Axis Options in Layouts

The axis line, ticks, and tick values of the “perpendicular” axis move
to the location indicated by the origin. The axis label is not moved.

Tip This option is often used to create Cartesian axes (axes centered at
ORIGIN=0).

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interactions This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

This option is ignored when the INCLUDERANGES= option is

specified.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0

prevents the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 885

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interactions This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

This option is ignored when the INCLUDERANGES= option is

specified.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0

prevents the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.
Axis Options for LAYOUT OVERLAY 919

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 885

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string values
that are displayed on the axis in place of the values in the TICKVALUELIST=
option. The strings map one-to-one positionally with the values that are listed in the
TICKVALUELIST= option.

Default The display of tick values is controlled by the

TICKVALUEFORMAT= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Interaction When this option is specified, the TICKVALUEFORMAT= option

is ignored.

Tip This option should be used with the TICKVALUELIST= option.

The number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness of a
collision-avoidance policy depends on the number of tick values, their length, and
the length of the axis. Which policies are valid depends on the axis on which this
option is used. For the Y and Y2 axes, the following policies are valid:
NONE
makes no attempt to avoid collisions between tick values. Tick values are display
even when they collide.
THIN
eliminates alternate tick values.
For the X and X2 axes, the following policies are valid:
ROTATE
rotates the tick values if a collision occurs. The TICKVALUEROTATION=
option specifies whether the values are rotated to a 45-degree diagonal or a 90-
degree vertical position. By default, the values are rotated to a 45-degree
diagonal position.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs. The
TICKVALUEROTATION= option specifies whether the values are rotated to a
45-degree diagonal or a 90-degree vertical position. By default, the values are
rotated to a 45-degree diagonal position.
ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
STAGGER
alternates the tick values between two rows.
920 Chapter 8 • Axis Options in Layouts

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.
THIN
eliminates alternate tick values.

Default THIN

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.
(format-options)
specifies one or more formatting options for major tick values. Together, these
options provide parameters for determining an optimal format (w.d, Ew.,
BESTw.) for displaying major tick values.
MAXWIDTH=integer
specifies the maximum width for displayed tick values. Values might be
rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values. Values
might be rounded or converted to E notation to fit into this width.

Default 6

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values. The actual number might vary based on other constraints.

Default 2

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it to
reduce the tick value width. The scale can be a named scale or a scientific-
notation scale. The EXTRACTSCALETYPE= option specifies the scale type.
The scale that is used is appended to the axis label, as shown in the following
example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the label
is truncated, and the scale is appended to the truncated label. Ellipses indicate
that the label was truncated, as shown in the following example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE
Axis Options for LAYOUT OVERLAY 921

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

The EXTRACTSCALE= suboption is ignored when the

INCLUDERANGES= option is specified.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis. In
that case, the tick value format might differ from the data
format even when a named format is applied to the data values.

See “boolean ” on page 1339 for other Boolean values that you
can use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.
DEFAULT
extracts a named scale. A named scale can be millions, billions, or
trillions for values of 999 trillion or less, or a multiple of 10 (denoted as
10^n) for values over 999 trillion. For large tick values, the scale factor is
set to ensure that the absolute value of the largest value is greater than 1.
For small fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.
SCIENTIFIC
extracts a scientific-notation scale. A scientific-notation scale is a
multiple of 10 expressed as 10^n for values greater than 1, or a multiple
of 1/10 expressed as 10^–n for values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.

DATA
uses the format that has been assigned to the column that is contributing to the
axis (or BEST6 if no format is assigned) in order to control the formatting of the
major tick values.
format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For details,
see Appendix 4, “SAS Formats Not Supported,” on page 1353.

Note If you specify a format that significantly reduces precision, because

of tick-value rounding, the plot data elements might not align
properly with the axis tick values. In that case, specify a tick-value
format with a higher precision.
922 Chapter 8 • Axis Options in Layouts

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

Interaction This option is ignored when the TICKDISPLAYLIST= option is

specified.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default An internal algorithm determines the tick marks, based on the actual
axis data range or the data range established by the VIEWMIN= and
VIEWMAX= options. By default, when this option is used, the only
tick values that appear are the tick values in numeric-list that fall
within the explicit data range (set by VIEWMIN= and VIEWMAX=)
or the implicit data range (set by the actual data minimum and data
maximum).

Restriction This option applies to linear axes only.

Requirement The tick values must be specified as a space-separated list of numeric

values, enclosed in parentheses.

Interactions This option overrides the INTEGER= option.

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option or the
DISPLAYSECONDARY= option does not display tick values.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then all
ticks in the tick list are displayed. This might result in some data not
being displayed. For example, data might not be displayed when the
VIEWMIN= value is greater than the actual data minimum, or when
the VIEWMAX= value is less than actual data maximum.

If TICKVALUEPRIORITY= TRUE, then the VIEWMIN= and

VIEWMAX= options are ignored if they are fully enclosed by the
numeric-list. The tick numeric-list can extend the implicit data range
of the axis, but cannot reduce it.

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option. If the minimum and maximum of the user-
specified values are within the data range, this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
Axis Options for LAYOUT OVERLAY 923

options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Restriction This option applies to linear axes only.

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored when the INCLUDERANGES= option is

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | VERTICAL
specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL rotates the tick values to a 45-degree diagonal position. The X

labels read left to right in a downward direction. The X2 labels
read left to right in an upward direction.
VERTICAL rotates the labels to a 90-degree vertical position. The labels are
always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for XAXISOPTS= and X2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESEQUENCE=(sequence-options)
specifies the tick values by start, end, and increment.
(sequence-options)
a space-separated list of the following name-value-pair options that control major
tick values. You must provide all three options.
START=number
specifies the value for the first tick mark.
END=number
specifies the value for the last tick mark.
INCREMENT=number
specifies the increment for intermediate tick marks between the first and last
tick marks. The END value always controls the last tick mark. The interval
between the last tick mark and the previous tick mark might not necessarily
be the INCREMENT value.
924 Chapter 8 • Axis Options in Layouts

Default An internal algorithm determines the tick marks, based on the actual
axis data range or the data range established by the VIEWMIN= and
VIEWMAX= options. By default, when this option is used, the only
tick values that appear are those that fall within the explicit data range
(set by VIEWMIN= and VIEWMAX=) or the implicit data range (set
by the actual data minimum and data maximum).

Interactions This option overrides the INTEGER= option.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the START= option value and the
VIEWMAX= option is set to the END= option value, then all ticks in
the tick sequence are displayed.

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display tick marks.

Tip The values in the sequence are formatted according to the setting for
the TICKVALUEFORMAT= option.

See TICKVALUELIST= option as an alternative for customizing tick

marks.

VIEWMAX=number
specifies the maximum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The maximum value in the data for the specified axis.

Interactions This option does not determine the maximum axis tick value that is
displayed. The THRESHOLDMAX= value is used to determine the
maximum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

This option is ignored when the INCLUDERANGES= option is

specified.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

VIEWMIN=number
specifies the minimum data value to include in the display. The value might be
adjusted by the threshold calculation.
Axis Options for LAYOUT OVERLAY 925

Default The minimum value in the data for the specified axis.

Interactions This option does not determine the minimum axis tick value that is
displayed. The THRESHOLDMIN= value is used to determine the
minimum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

This option is ignored when the INCLUDERANGES= option is

specified.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Options for Log Axes Only

The options that are documented in this section can be used with the LOGOPTS= axis
option. The following table provides a summary of the options.

Log Axis Option Description

BASE Specifies the base of the logarithmic scale for

the axis values.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether minor ticks are displayed.

THRESHOLDMAX Specifies a bias for including one more tick

mark at the maximum end of the axis.

THRESHOLDMIN Specifies a bias for including one more tick

mark at the minimum end of the axis.

TICKINTERVALSTYLE Specifies how to scale and format the values

for major tick marks.

TICKVALUEFORMAT= Specifies how to format the values for major

tick marks.
926 Chapter 8 • Axis Options in Layouts

Log Axis Option Description

TICKVALUELIST Specifies the tick values for a log axis as a

space-separated list.

TICKVALUEPRIORITY Specifies whether the TICKVALUELIST=

specification can extend the axis data range.

VALUESTYPE Specifies the scale that the system uses when

interpreting the TICKVALUELIST=,
VIEWMAX=, and VIEWMIN= option values.

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

BASE=10 | 2 | E
specifies the base of the logarithmic scale for the axis values.

Default 10

Restriction This option applies to log axes only.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.
Axis Options for LAYOUT OVERLAY 927

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
The following figure shows the minor grid lines set to light blue, dotted lines on a
base-10 log axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Default Eight ticks with nine intervals (BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT. ).

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.
928 Chapter 8 • Axis Options in Layouts

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to log axes only.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

See “Adjusting Axis Thresholds” on page 885

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to log axes only.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.
Axis Options for LAYOUT OVERLAY 929

See “Adjusting Axis Thresholds” on page 885

TICKINTERVALSTYLE=AUTO | LOGEXPAND | LOGEXPONENT | LINEAR

specifies how to scale and format the values for major tick marks.
AUTO
selects a LOGEXPAND, LOGEXPONENT, or LINEAR representation
automatically based on the range of the data. When the data range is small
(within an order of magnitude), a LINEAR representation is typically used. Data
ranges that encompass several orders of magnitude typically use the
LOGEXPAND or LOGEXPONENT representation.
LOGEXPAND
places the major tick marks at uniform intervals at integer powers of the base.
The tick values are expanded as follows:

Base=10

Base=2

Base=E

LOGEXPONENT
places the major tick marks at uniform intervals at integer powers of the base.
The tick values are only the integer exponents for all bases.

LINEAR
places the major tick marks at non-uniform intervals that cover the range of the
data.

Default AUTO

Restrictions This option applies to log axes only.

For LOGEXPONENT, formats on data columns contributing to the

axis are ignored. For LOGEXPAND, formats on data columns
contributing to the axis are ignored, although any "named format" on
the column is retained. For LINEAR, ticks values are automatically
formatted when the column format is not assigned or one of w.d, Ew.,
or BESTw. Other formats (SAS defined or user-defined) are used if
specified.

GTL currently honors most but not every SAS format. For details, see
Appendix 4, “SAS Formats Not Supported,” on page 1353.
930 Chapter 8 • Axis Options in Layouts

Note When BASE=10 and LOGEXPAND or LOGEXPONENT is used, an

intermediate tick is added whenever the axis data range is less than or
equal to 1.5 powers of 10.

Tip If you use TICKINTERVALSTYLE=LOGEXPONENT, then you

might want to include information in the axis label about which base
is used.

TICKVALUEFORMAT=DATA | format
specifies how to format the values for major tick marks.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
DATA
uses the format that has been assigned to the column that is contributing to the
axis (or BEST6 if no format is assigned) in order to control the formatting of the
major tick values.
format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For details,
see Appendix 4, “SAS Formats Not Supported,” on page 1353.

Restriction This option applies to log axes only.

Interactions This option is ignored when

TICKINTERVALSTYLE=LOGEXPONENT.

When TICKINTERVALSTYLE=LOGEXPAND, this option is

honored for the base 10 and base 2 logarithmic scales, and is ignored
for the base E scale.

When TICKINTERVALSTYLE=LINEAR, this option is honored for

the base 10, base 2, and base E logarithmic scales.

See BASE=

TICKINTERVALSTYLE=

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default Only the tick values specified in the list that fall within the explicit
data range set by the VIEWMIN= and VIEWMAX= options or by
the implicit data range set by the actual data minimum and data
maximum are displayed. An internal algorithm determines the tick
marks.

Requirements The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

The values that you specify must be appropriate for the

VALUESTYPE= specification. Otherwise, unexpected results might
occur. If VALUESTYPE=EXPANDED is in effect (default), specify
increments of the log base power such as 0.1, 1, 10, 100, and so on,
on a base 10 log axis, for example. If VALUESTYPE=EXPONENT
Axis Options for LAYOUT OVERLAY 931

is in effect, specify integer increments of the log base power

exponent such as 1, 2, 3, and so on.

Interactions The VALUESTYPE= option determines how the values in the list
are interpreted.

The VIEWMIN= and VIEWMAX= options alter the axis data

range. If the VIEWMIN= option is set to the minimum tick list
value and the VIEWMAX= option is set to the maximum tick list
value, then all ticks in the tick list are displayed. This might result in
some data not being displayed. For example, data might not be
displayed when the VIEWMIN= value is greater than the actual data
minimum, or when the VIEWMAX= value is less than actual data
maximum.

If the VIEWMIN= value is greater than the actual data minimum or

the VIEWMAX= value is less than actual data maximum, some data
might not be displayed.

This option is ignored if the DISPLAY= or the

DISPLAYSECONDARY= option does not display the tick values.

See VIEWMIN= and VIEWMAX= options for controlling the data

range

TICKINTERVALSTYLE= for specifying the scale and format of the

major tick values

TICKVALUEPRIORITY= for controlling the behavior of the

TICKVALUELIST= option

BASE= for specifying the log base

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by the TICKVALUELIST= option. If the
minimum and maximum of the user-specified values are within the data range,
this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.
932 Chapter 8 • Axis Options in Layouts

This option is ignored if the TICKVALUELIST= option is not

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

VALUESTYPE=EXPANDED | EXPONENT
specifies the scale that the system uses when interpreting the TICKVALUELIST=,
VIEWMAX=, and VIEWMIN= option values. Use this option to choose your
preferred way of specifying log-axis values.

EXPANDED values are interpreted as powers of the base such as 0.1, 1, 10,
100, and so on, for base 10, for example.
EXPONENT values are interpreted as integer exponents of the base such as 1,
2, 3, and so on, for base 10, base 2, and base E.

Default EXPANDED

Note This option does not change the style of the tick values that are
displayed on the axis. It changes only how the VIEWMIN=,
VIEWMAX=, and TICKVALUELIST= option values are interpreted by
the system.

Tip This option is particularly useful when BASE=E.

Examples The following example specifies VIEWMIN= and VIEWMAX= as

exponent values instead of as expanded values on an expanded Base 10
log axis. This results in X-axis tick values of 10, 100, 1000, 10000, and
100000.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexpand
valuestype=exponent
viewmin=1 viewmax=5));

The following example specifies TICKVALUELIST= as a list of

expanded values instead of exponent values on an exponent Base 10 log
axis. This results in X-axis tick values of 1, 2, 3, 4, and 5.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexponent
tickvaluepriority=true
valuestype=expanded
tickvaluelist=(10 100 1000 10000 100000)));

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Requirement The value that you specify must be appropriate for the
VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is in
effect (default), specify an increment of the log base power such as
Axis Options for LAYOUT OVERLAY 933

0.1, 1, 10, 100, and so on, on a base 10 log axis, for example. If
VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

If an invalid value is specified for the VIEWMAX= option, the

default value for VIEWMAX= is used instead. In that case, if the
default value for VIEWMAX= is less than the value specified by the
VIEWMIN= option, then the VIEWMIN= and VIEWMAX= values
are swapped.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to 1.5
powers of 10.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Examples The following example specifies a value of 100,000 as an expanded

value on a base 10 log axis:
VIEWMAX=100000

The following example specifies a value of 100,000 as an exponent

value on a base 10 log axis:
VIEWMAX=5

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Requirement The value that you specify must be appropriate for the
VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is in
effect (default), specify an increment of the log base power such as
0.1, 1, 10, 100, and so on, on a base 10 log axis, for example. If
VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is interpreted.

934 Chapter 8 • Axis Options in Layouts

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to 1.5
powers of 10.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Examples The following example specifies a value of 0.1 as an expanded value

on a base 10 log axis:
VIEWMIN=0.1

The following example specifies a value of 0.1 as an exponent value

on a base 10 log axis:
VIEWMIN=-1

Options for Time Axes Only

The options that are documented in this section can be used with the TIMEOPTS= axis
option. The following table provides a summary of the options.

Time Axis Option Description

INCLUDERANGES on page 935 Specifies one or more ranges for a broken

axis.

INTERVAL Specifies the time interval between major tick

marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time

interval that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKINTERVAL Specifies the time interval between minor

ticks.

MINORTICKS Specifies whether minor ticks are displayed.

SPLITTICKVALUE Specifies whether to split the tick values on an

X or X2 axis, if possible. This option is not
available on a Y or Y2 axis.
Axis Options for LAYOUT OVERLAY 935

Time Axis Option Description

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision on an X or X2 axis. This option is
not available on a Y or Y2 axis.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the order of the tick values for a

time axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification

can extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on

the X and X2 axes.

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

INCLUDERANGES=(start–end <start2–end2 startN–endN …>)

specifies the ranges for a broken axis.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
start
specifies a SAS time, date, or date-time constant, or the keyword MIN. MIN
specifies the minimum data value.
end
specifies a SAS time, date, or date-time constant, or the keyword MAX. MAX
specifies the maximum data value.
936 Chapter 8 • Axis Options in Layouts

The following figure shows a time axis, broken into ranges '01Jan2001'd–'01May2003'd
and '01Jan2005'd–'01Oct2005'd.

As shown in the figure, break lines are drawn to indicate the break in the axis.

Restrictions This option is valid for linear and time axes in an OVERLAY layout only.

Only one axis can be broken. If this option is specified for both axes, then it is
honored for the vertical axis and ignored for the horizontal axis.

When plots are associated with the X and X2 axes or with Y and Y2 axes,
neither axis can be broken.

A binned heat map or histogram axis cannot be broken.

A broken axis is not supported in vector graphics output. When a broken axis
is specified and vector graphics output is requested, the graph is converted
into an image instead. A note indicating the conversion is written to the SAS
log. To restore the vector graphics output in that case, remove the
INCLUDERANGES= option from the LINEAROPTS= or TIMEOPTS=
option.

Requirements All of the ranges must be enclosed in parenthesis.

You must specify each range as a starting value, a hyphen, and an ending
value. You must separate adjacent ranges with a space.

Each range must be nonzero. A zero range such as 12–12 is considered

invalid.

Interactions When this option is specified, axis options THRESHOLDMIN=,

THRESHOLDMAX=, VIEWMIN=, VIEWMAX=, and
TICKVALUEPRIORITY= are ignored. Suboption EXTRACTSCALE= of the
TICKVALUEFORMAT= option is also ignored.

When this option is specified, the plot statement TIP= and URL= options are
ignored.

Notes When this option is specified, data-clipping might occur for the following:
plot markers and marker characters, box-plot outlier markers, fixed-position
Axis Options for LAYOUT OVERLAY 937

data labels, needle plots and fringe plots in the X direction, reference lines and
drop lines on the broken axis, axis tables, and relative bubble plots.

Curve label positions are based on the contiguous axis data range. When curve
labels are specified with a broken axis, the curve label positions might not be
ideal.

Tip Starting with the third maintenance release of SAS 9.4, you can use the
AXISBREAKTYPE= and AXISBREAKSYMBOL= options in the
BEGINGRAPH statement to display the break in the axis as only a symbol on
the axis line.

See “Creating a Broken Time Axis” in SAS Graph Template Language: User's
Guide

Example includeranges=('01Jan2001'd-'01May2003'd '01Jan2005'd-'01Oct2005'd)

INTERVAL=interval
specifies the time interval between major ticks. Valid interval keywords are AUTO,
SECOND, MINUTE, HOUR, DAY, TENDAY, WEEK, SEMIMONTH, MONTH,
QUARTER, SEMIYEAR, YEAR.
Table 8.1 Time Intervals

Default tick value

INTERVAL Unit Tick interval format

AUTO DATE, TIME, or automatically automatically

DATETIME chosen chosen

SECOND TIME or second TIME8.

DATETIME

MINUTE TIME or minute TIME8.

DATETIME

HOUR TIME or hour TIME8.

DATETIME

DAY DATE or day DATE9.

DATETIME

TENDAY DATE or 10 days DATE9.

DATETIME

WEEK DATE or 7 days DATE9.

DATETIME

SEMIMONTH DATE or 1st and 16th of each DATE9.

DATETIME month

MONTH DATE or month MONYY7.

DATETIME

QUARTER DATE or 3 months YYQC6.

DATETIME
938 Chapter 8 • Axis Options in Layouts

Default tick value

INTERVAL Unit Tick interval format

SEMIYEAR DATE or 6 months MONYY7.

DATETIME

YEAR DATE or year YEAR4.

DATETIME

Default AUTO. An appropriate interval is chosen based on the data and the
column date, date-time, or time format.

Restriction This option applies to time axes only.

Requirement The data column(s) mapped to a time axis must be in the same
duration units: TIME, DATE, or DATETIME. The selection of an
interval must be consistent with the duration unit. For example, if the
data are in time units, you can specify only AUTO, SECOND,
MINUTE, HOUR.

Interaction This option is ignored if the TICKVALUELIST= option is used.

INTERVALMULTIPLIER=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the axis.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default 1

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= option is used.

Tip Use the INTERVAL= option to specify a different time interval.

Examples To specify 3-month intervals:

INTERVAL=MONTH INTERVALMULTIPLIER=3

To specify 10-year intervals:

INTERVAL=YEAR INTERVALMULTIPLIER=10

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.
Axis Options for LAYOUT OVERLAY 939

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
The following figure shows the minor grid lines set to light blue, dotted lines on a time
axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.
940 Chapter 8 • Axis Options in Layouts

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKINTERVAL=interval
specifies the time interval between minor ticks. See Table 8.1 on page 937 for
information about the intervals that you can select. The interval that you select must
be consistent with the axis data duration units such as TIME, DATE, or DATETIME.
For example, if the axis data is in TIME units, then you must specify AUTO,
SECOND, MINUTE, or HOUR.

Default AUTO

Interactions This option is ignored if the TICKVALUELIST= option is used.

This option is ignored if the MINORTICKINTERVAL= setting is

greater than the INTERVAL= setting.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Interactions The number of minor ticks is dependent on the value of the

MINORTICKINTERVAL= option, if specified. If
MINORTICKINTERVAL= is not specified, then it is dependent on the
value of the INTERVAL= option.

This option is ignored if the TICKVALUELIST= option is used, or if

the DISPLAY= or DISPLAYSECONDARY= option does not display
the tick marks.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

SPLITTICKVALUE=TRUE | FALSE
specifies whether to split the tick values on an X or X2 axis, if possible. This option
is not available for a Y or Y2 axis.
Axis Options for LAYOUT OVERLAY 941

TRUE
splits the axis tick values into two lines allowing more tick values to appear. For
example, with INTERVAL= MONTH, this is how tick values are split:

FALSE
does not split the axis tick values. For example, when this option specifies
FALSE, this is how the tick values in the previous example appear:

Typically, fewer tick values fit, causing thinning, rotation, or staggering of the
values. See the TICKVALUEFITPOLICY= option.

Default TRUE

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUEFORMAT= option is used.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an X or X2 axis. This option is
not available for the Y and Y2 axes. The effectiveness of a collision-avoidance
policy depends on the number of tick values, their length, and the length of the axis.
The following policies are valid:
NONE
makes no attempt to avoid collisions between tick values. Tick values are display
even when they collide.
ROTATE
rotates the tick values if a collision occurs. The TICKVALUEROTATION=
option specifies whether the values are rotated to a 45-degree diagonal or a 90-
degree vertical position. By default, the values are rotated to a 45-degree
diagonal position.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs. The
TICKVALUEROTATION= option specifies whether the values are rotated to a
45-degree diagonal or a 90-degree vertical position. By default, the values are
rotated to a 45-degree diagonal position.
ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
STAGGER
alternates the tick values between two rows.
STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
942 Chapter 8 • Axis Options in Layouts

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.
THIN
eliminates alternate tick values.

Default THIN

Restriction This option is valid only for the X and X2 axes.

Interaction When SPLITTICKVALUE= TRUE, this option is ignored and only the
THIN policy is used.

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format | DATA
specifies how to format the values for major tick marks.
format
specifies a SAS date, time, or datetime format to control how the major tick
values are displayed. This format must be in the same duration units as the data
column(s) mapped to a time axis: TIME, DATE, or DATETIME and should be
appropriate for the value of the INTERVAL= option. For example, if
INTERVAL=MONTH and there are two years of data displayed on the axis, then
choosing TICKVALUEFORMAT=YEAR. would result in several ticks having
the same year value.
DATA
specifies that the SAS date, time, or datetime format associated with the data
column assigned to the axis be used to control how the major tick values are
displayed.

Default The default format used by the INTERVAL= option. The default does
not apply if TICKVALUELIST= is specified.

Restrictions This option applies to time axes only.

GTL currently honors most but not every SAS format. For details, see
Appendix 4, “SAS Formats Not Supported,” on page 1353.

TICKVALUELIST=(time-constant-list | date-constant-list | datetime-constant-list |

numeric-list)
specifies the tick values for a time axis as list.

Default An internal algorithm determines the tick values.

Restrictions This option applies to time axes only.

If TICKVALUEPRIORITY= is set to FALSE, then this option does

not extend the data range of the axis. If the values fall within the
default data range or that specified by the VIEWMIN= or
VIEWMAX= options, then they are used.

Requirement The tick values must be specified as a space-separated list of values

enclosed in parentheses. The items in the list must be in the same
duration units as the data mapped to the axis: TIME, DATE, or
DATETIME. The values can be expressed as SAS TIME, DATE, or
Axis Options for LAYOUT OVERLAY 943

DATETIME constants (for example, "13:23"T, "11MAY06"D, or

"11MAY06:13:23"DT) or their numeric equivalents.

Interactions The values in the list are formatted according to the format specified
on the TICKVALUEFORMAT= option. If TICKVALUEFORMAT=
is not used, then the values are formatted according to the column
format (the default TICKVALUEFORMAT value is not applied to
these values).

If this option is specified, the SPLITTICKVALUE= and

INTERVAL= options are ignored.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by the TICKVALUELIST= option. If the
minimum and maximum of the user-specified values are within the data range,
this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

This option is ignored when the INCLUDERANGES= option is

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | VERTICAL
specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL rotates the tick values to a 45-degree diagonal position. The X

labels read left to right in a downward direction. The X2 labels
read left to right in an upward direction.
VERTICAL rotates the labels to a 90-degree vertical position. The labels are
always drawn from bottom to top.

Default DIAGONAL
944 Chapter 8 • Axis Options in Layouts

Restriction This option is valid for XAXISOPTS= and X2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

This option is ignored when the INCLUDERANGES= option is

specified.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

This option is ignored when the INCLUDERANGES= option is

specified.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Details
The LAYOUT OVERLAY statement provides the XAXISOPTS=, YAXISOPTS=,
X2AXISOPTS=, Y2AXISOPTS= options that enable you to manage the axis display
separately for the X, Y, X2, and Y2 axes. The following example template uses the
YAXISOPTS= option to manage the grid lines, tick marks, and tick values on a Y axis:
Axis Options for LAYOUT OVERLAY3D 945

begingraph;
layout overlay /
yaxisopts=(
griddisplay=on
display=(ticks tickvalues)
);
seriesplot x=month y=predict;
endlayout;
endgraph;

Within an OVERLAY layout block, each plot axis is always of a particular type. In the
default cases, the axis type is always DISCRETE, LINEAR, or TIME. The TYPE=
option enables you to specify an axis type that overrides the default. For example, when
appropriate for the data, you can request a LOG axis. When you override the default axis
type, you must be sure to specify the correct axis type for the plot(s) that you are
defining.
Each axis type has features specific to that type, and the following axis options enable
you to specify features for the different types: DISCRETEOPTS= , LINEAROPTS= ,
LOGOPTS= , and TIMEOPTS= . One or more of these options can be specified for an
axis, but the specified settings are applied only to the axis type that supports them.

Axis Options for LAYOUT OVERLAY3D

Axis options for the plots within an OVERLAY3D layout.
Note: Unless otherwise indicated in an option description, each axis option is available for
the X, Y, and Z axis.
See: “LAYOUT OVERLAY3D Statement” on page 152

Syntax
Axis options for the plots within an OVERLAY3D layout are specified with the
following options on a LAYOUT OVERLAY3D statement:
XAXISOPTS=(axis-options)
YAXISOPTS=(axis-options)
ZAXISOPTS=(axis-options)

General Options for All Axes in an Overlay3D

The options that are documented in this section can be used with either axis type that is
supported within an OVERLAY3D layout. Subsequent sections in the chapter document
the axis options that are available only for the specific axis type: linear or time.

Statement Option Description

DISPLAY Controls which axis features are displayed.

GRIDATTRS Specifies the attributes of the grid lines.

GRIDDISPLAY Specifies whether axis grid lines are

displayed.
946 Chapter 8 • Axis Options in Layouts

Statement Option Description

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the

axis label.

LINEAROPTS Specifies options for a standard numeric

interval axis.

OFFSETMAX Reserves an area at the maximum end of the

axis. No tick marks are displayed in the
reserved area.

OFFSETMIN Reserves an area at the minimum end of the

axis. No tick marks are displayed in the
reserved area.

TICKVALUEATTRS Specifies the color and font attributes of the

axis tick values.

TIMEOPTS Specifies options for a TIME axis.

TYPE Specifies the type of axis to use.

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on the primary axis.
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed
NONE
specifies that no axis features are displayed
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL displays the axis label

LINE displays the axis line
TICKS displays the tick marks
TICKVALUES displays the values that are represented by the major tick
marks

Default STANDARD

Tips The default line attributes for the axis line and axis tick marks are defined
in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.
Axis Options for LAYOUT OVERLAY3D 947

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall chart
can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template to
absolutely control the display of grid lines or to allow interaction with the current
style to decide whether grid lines are displayed.
AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element in
the current style contains DisplayOpts="ON."
AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in the
current style contains DisplayOpts="OFF."
ON
specifies that grid lines are always displayed. The current style has no override.
OFF
specifies that grid lines are never displayed. The current style has no override.
The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines style
element. Most supplied templates use the default setting AUTO_OFF to indicate a
preference for not displaying grid lines, but allowing the style to override.

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no
948 Chapter 8 • Axis Options in Layouts

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

ON any value yes

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style has
no preference about grid lines and the graphics template setting for grid
lines is always used.

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The list
form implies that all included string literals or dynamics will be concatenated.

Default The default label is derived from the primary plot in the layout. For
more information, see “When Plots Share Data and a Common Axis”
on page 880.

Interaction This option is ignored if the DISPLAY= option does not display the
axis label.

Note If the axis label is too long to fit along the axis, then it is truncated by
default.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= option does not display the
axis label.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LINEAROPTS=(linear-axis-options)
specifies one or more options for a numeric interval axis. Options must be enclosed
in parentheses. Each option is specified as a name = value pair and each pair is space
separated.

Interaction This option is ignored if the axis type is not LINEAR.

See “Options for Linear Axes Only” on page 950 for the options that you
can use for linear-axis-options.

OFFSETMAX=AUTO | number
reserves an area at the maximum end of the axis. No tick marks are displayed in the
reserved area.
Axis Options for LAYOUT OVERLAY3D 949

AUTO
reserves just enough area to fully display markers and other graphical features
near the maximum end of an axis.
number
specifies the offset as a decimal proportion of the full axis length. For a
continuous axis, the offset follows the highest data value or highest tick value,
whichever is greater.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886

OFFSETMIN=AUTO | number
reserves an area at the minimum end of the axis. No tick marks are displayed in the
reserved area.
AUTO
reserves just enough area to fully display markers and other graphical features
near the minimum end of an axis.
number
specifies the offset as a decimal proportion of the full axis length. For a
continuous axis, the offset precedes the lowest data value or lowest tick value,
whichever is less.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= option does not display tick
values.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TIMEOPTS=(time-axis-options)
specifies one or more options for a time axis.

Requirements Columns associated with a time axis must be in SAS time, SAS
date, or SAS datetime units and have an associated SAS time, date,
or datetime format.

Options must be enclosed in parentheses. Each option is specified as

a name = value pair and each pair is space separated.
950 Chapter 8 • Axis Options in Layouts

Interaction This option is ignored if the axis type is not TIME.

See “Options for Time Axes Only” on page 958 for the options that
you can use for time-axis-options.

TYPE=AUTO | LINEAR | TIME

specifies the type of axis to use.
AUTO
requests that the axis type be automatically determined by the plot or the overlay
contents.
LINEAR
uses a LINEAR axis if possible. You can add a LINEAROPTS= option list to
customize this axis type.
TIME
uses a TIME axis if possible. Data for this axis must be SAS time, SAS date, or
SAS datetime values. You can add a TIMEOPS= option list to customize this
axis type.

Default AUTO

Interactions If this option is set to anything other than AUTO, then plots within the
layout are dropped from the display if their data types or data ranges
do not match the axis type requirements. For more information, see
“Plot Axis Types Must Agree on Common Axes” on page 883.

After the axis type is determined (whether you set a specific type or
AUTO is in effect), only options supported by that axis type can be
used. For example, if TYPE=TIME, then only the general
OVERLAY3D axis options and those available on TIMEOPS= are
supported.

Options for Linear Axes Only

This section documents the options that can be used with the LINEAROPTS= axis
option. The following table provides a summary of the options.

Linear Axis Option Description

INTEGER Specifies that evenly spaced integer values are

used for tick marks.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether minor ticks are displayed.

THRESHOLDMAX Specifies a bias for including one more tick

mark at the maximum end of the axis.
Axis Options for LAYOUT OVERLAY3D 951

Linear Axis Option Description

THRESHOLDMIN Specifies a bias for including one more tick

mark at the minimum end of the axis.

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the
TICKVALUELIST= option.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the order of the tick values for a

linear axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification

(TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the
axis data range.

TICKVALUESEQUENCE Specifies the tick values for a linear axis by

start, end, and increment.

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the
TICKVALUEFORMAT= option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAM or HISTOGRAMPARM statement.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.
952 Chapter 8 • Axis Options in Layouts

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
The following figure shows the minor grid lines set to light blue, dotted lines on a linear
axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.
Axis Options for LAYOUT OVERLAY3D 953

Defaults Four ticks with five intervals in the first maintenance release of SAS
9.4 and earlier releases.

One tick with two intervals starting with the second maintenance
release of SAS 9.4.

Interactions The DISPLAY= option specification must include TICKS for this
option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Interaction This option is ignored if the TICKVALUELIST= option is used or if

the tick marks are not enabled by the DISPLAY= option.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.
954 Chapter 8 • Axis Options in Layouts

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 885

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 885

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string values
that are displayed on the axis in place of the values in the TICKVALUELIST=
option. The strings map one-to-one positionally with the values that are listed in the
TICKVALUELIST= option.

Default The display of tick values is controlled by the

TICKVALUEFORMAT= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Interaction When this option is specified, the TICKVALUEFORMAT= option

is ignored.

Tip This option should be used with the TICKVALUELIST= option.

The number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.
Axis Options for LAYOUT OVERLAY3D 955

(format-options)
specifies one or more formatting options for major tick values. Together, these
options provide parameters for determining an optimal format (w.d, Ew.,
BESTw.) for displaying major tick values.
MAXWIDTH=integer
specifies the maximum width for displayed tick values. Values might be
rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values. Values
might be rounded or converted to E notation to fit into this width.

Default 6

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values. The actual number might vary based on other constraints.

Default 2

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it to
reduce the tick value width. The scale can be a named scale or a scientific-
notation scale. The EXTRACTSCALETYPE= option specifies the scale type.
The scale that is used is appended to the axis label, as shown in the following
example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the label
is truncated, and the scale is appended to the truncated label. Ellipses indicate
that the label was truncated, as shown in the following example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis. In
that case, the tick value format might differ from the data
format even when a named format is applied to the data values.
956 Chapter 8 • Axis Options in Layouts

See “boolean ” on page 1339 for other Boolean values that you
can use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.
DEFAULT
extracts a named scale. A named scale can be millions, billions, or
trillions for values of 999 trillion or less, or a multiple of 10 (denoted as
10^n) for values over 999 trillion. For large tick values, the scale factor is
set to ensure that the absolute value of the largest value is greater than 1.
For small fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.
SCIENTIFIC
extracts a scientific-notation scale. A scientific-notation scale is a
multiple of 10 expressed as 10^n for values greater than 1, or a multiple
of 1/10 expressed as 10^–n for values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.

DATA
uses the format that has been assigned to the column that is contributing to the
axis (or BEST6 if no format is assigned) in order to control the formatting of the
major tick values.
format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For details,
see Appendix 4, “SAS Formats Not Supported,” on page 1353.

Note If you specify a format that significantly reduces precision, because

of tick-value rounding, the plot data elements might not align
properly with the axis tick values. In that case, specify a tick-value
format with a higher precision.

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

Interaction This option is ignored when the TICKDISPLAYLIST= option is

specified.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default An internal algorithm determines the tick values.

Restriction This option applies to linear axes only.

Requirement The tick values must be specified as a space-separated list of numeric

values, enclosed in parentheses.

Interactions This option overrides the INTEGER= option.

Axis Options for LAYOUT OVERLAY3D 957

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option does not display tick values.

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option. If the minimum and maximum of the user-
specified values are within the data range, this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the actual data minimum and data
maximum.

Default FALSE

Restriction This option applies to linear axes only.

Interactions This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option does not display the
tick values.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUESEQUENCE=(sequence-options)
specifies the tick values by start, end, and increment.
(sequence-options)
a space-separated list of the following name-value-pair options that control major
tick values. You must provide all three options.
START=number
specifies the value for the first tick mark.
END=number
specifies the value for the last tick mark.
INCREMENT=number
specifies the increment for intermediate tick marks between the first and last
tick marks. The END value always controls the last tick mark. The interval
between the last tick mark and the previous tick mark might not necessarily
be the INCREMENT value.

Default An internal algorithm determines the tick marks.

Interactions This option overrides the INTEGER= option.

958 Chapter 8 • Axis Options in Layouts

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the DISPLAY= option does not display tick
marks.

Tip The values in the sequence are formatted according to the setting for
the TICKVALUEFORMAT= option.

See TICKVALUELIST= option as an alternative for customizing tick

marks.

Options for Time Axes Only

This section documents the options that can be used with the TIMEOPTS= axis option.
The following table provides a summary of the options.

Time Axis Option Description

INTERVAL Specifies the time interval between major tick

marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time

interval that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKINTERVAL Specifies the time interval between minor

ticks.

MINORTICKS Specifies whether minor ticks are displayed.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the order of the tick values for a

time axis as list.

INTERVAL=interval
specifies the time interval between major ticks. Valid interval keywords are AUTO,
SECOND, MINUTE, HOUR, DAY, TENDAY, WEEK, SEMIMONTH, MONTH,
QUARTER, SEMIYEAR, YEAR.
Table 8.2 Time Intervals

Default tick value

INTERVAL Unit Tick interval format

AUTO DATE, TIME, or automatically automatically

DATETIME chosen chosen
Axis Options for LAYOUT OVERLAY3D 959

Default tick value

INTERVAL Unit Tick interval format

SECOND TIME or second TIME8.

DATETIME

MINUTE TIME or minute TIME8.

DATETIME

HOUR TIME or hour TIME8.

DATETIME

DAY DATE or day DATE9.

DATETIME

TENDAY DATE or 10 days DATE9.

DATETIME

WEEK DATE or 7 days DATE9.

DATETIME

SEMIMONTH DATE or 1st and 16th of each DATE9.

DATETIME month

MONTH DATE or month MONYY7.

DATETIME

QUARTER DATE or 3 months YYQC6.

DATETIME

SEMIYEAR DATE or 6 months MONYY7.

DATETIME

YEAR DATE or year YEAR4.

DATETIME

Default AUTO. An appropriate interval is chosen based on the data and the
column date, date-time, or time format.

Restriction This option applies to time axes only.

Requirement The data column(s) mapped to a time axis must be in the same
duration units: TIME, DATE, or DATETIME. The selection of an
interval must be consistent with the duration unit. For example, if the
data are in time units, you can specify only AUTO, SECOND,
MINUTE, HOUR.

Interaction This option is ignored if the TICKVALUELIST= option is used.

INTERVALMULTIPLIER=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the axis.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
960 Chapter 8 • Axis Options in Layouts

Default 1

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= option is used.

Tip Use the INTERVAL= option to specify a different time interval.

Examples To specify 3-month intervals:

INTERVAL=MONTH INTERVALMULTIPLIER=3

To specify 10-year intervals:

INTERVAL=YEAR INTERVALMULTIPLIER=10

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
Axis Options for LAYOUT OVERLAY3D 961

The following figure shows the minor grid lines set to light blue, dotted lines on a time
axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKINTERVAL=interval
specifies the time interval between minor ticks. See Table 8.2 on page 958 for
information about the intervals that you can select. The interval that you select must
be consistent with the axis data duration units such as TIME, DATE, or DATETIME.
For example, if the axis data is in TIME units, then you must specify AUTO,
SECOND, MINUTE, or HOUR.

Default AUTO

Interactions This option is ignored if the TICKVALUELIST= option is used.

This option is ignored if the MINORTICKINTERVAL= setting is

greater than the INTERVAL= setting.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.
962 Chapter 8 • Axis Options in Layouts

Default FALSE

Interactions The number of minor ticks is dependent on the value of the

MINORTICKINTERVAL= option, if specified. If
MINORTICKINTERVAL= is not specified, then it is dependent on the
value of the INTERVAL= option.

This option is ignored if the TICKVALUELIST= option is used or if

the tick marks are not enabled by the DISPLAY= option.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEFORMAT=format | DATA
specifies how to format the values for major tick marks.
format
specifies a SAS date, time, or datetime format to control how the major tick
values are displayed. This format must be in the same duration units as the data
column(s) mapped to a time axis: TIME, DATE, or DATETIME and should be
appropriate for the value of the INTERVAL= option. For example, if
INTERVAL=MONTH and there are two years of data displayed on the axis, then
choosing TICKVALUEFORMAT=YEAR. would result in several ticks having
the same year value.
DATA
specifies that the SAS date, time, or datetime format associated with the data
column assigned to the axis be used to control how the major tick values are
displayed.

Default The default format used by the INTERVAL= option. The default does
not apply if TICKVALUELIST= is specified.

Restrictions This option applies to time axes only.

GTL currently honors most but not every SAS format. For details, see
Appendix 4, “SAS Formats Not Supported,” on page 1353.

TICKVALUELIST=(time-constant-list | date-constant-list | datetime-constant-list |

numeric-list)
specifies the tick values for a time axis as list.

Default An internal algorithm determines the tick values.

Restrictions This option applies to time axes only.

If TICKVALUEPRIORITY= is set to FALSE, then this option does

not extend the data range of the axis. If the values fall within the
default data range, then they are used.

Requirement The tick values must be specified as a space-separated list of values

enclosed in parentheses. The items in the list must be in the same
duration units as the data mapped to the axis: TIME, DATE, or
DATETIME. The values can be expressed as SAS TIME, DATE, or
DATETIME constants (for example, "13:23"T, "11MAY06"D, or
"11MAY06:13:23"DT) or their numeric equivalents.
Axis Options for LAYOUT LATTICE 963

Interactions The values in the list are formatted according to the format specified
on the TICKVALUEFORMAT= option. If TICKVALUEFORMAT=
is not used, then the values are formatted according to the column
format (the default TICKVALUEFORMAT value is not applied to
these values).

If this option is specified, then the INTERVAL= option is ignored.

Details
The LAYOUT OVERLAY3D statement provides XAXISOPTS=, YAXISOPTS=, and
ZAXISOPTS= options that enable you to manage the axis display separately for the X,
Y, and Z axes. The following example template uses the YAXISOPTS= option to
manage the display of grid lines, tick marks, and tick values on a Y axis:
begingraph;
layout overlay3d /
yaxisopts=(
griddisplay=on
display=(ticks tickvalues)
);
bihistogram3dparm x=rater y=customer z=percent;
endlayout;
endgraph;

Within an OVERLAY3D layout block, each plot axis is always either a linear or a time
axis. The default axis-type setting is AUTO, which specifies that the axis type be
automatically determined by the plot or the overlay contents. The TYPE= option enables
you to explicitly specify either a LINEAR or a TIME axis type. When you override the
default axis type, you must be sure to specify the correct axis type for the plot(s) that you
are defining.
Each axis type has features specific to that type, and the axis options LINEAROPTS=
and TIMEOPS= enable you to specify features for a linear or a time axis. You can
combine one or more general axis options with the options for the specific axis type.
However, specified settings are applied only to the axis type that supports them. For
example, if you specify general axis options with time axis options and the generated
graph produces a linear axis type, then the time axis options are ignored.

Axis Options for LAYOUT LATTICE

Axis options for the plots within a LATTICE layout.
See: “LAYOUT LATTICE Statement” on page 111

Syntax
Axis options for the plots within a LATTICE layout are specified with the following
statements within a LAYOUT LATTICE statement block:
COLUMNAXIS /external-axis-options
ROWAXIS /external-axis-options
964 Chapter 8 • Axis Options in Layouts

General Options for All Axes in a Lattice

The options that are documented in this section can be used with any of the axis types
that are supported in a LAYOUT LATTICE statement. Subsequent sections in the
chapter document the axis options that are available only for specific axis types: discrete,
linear, log, or time axes.

Statement Option Description

DISCRETEOPTS Specifies features for a discrete axis.

DISPLAY Controls which axis features are displayed on

the primary axis.

DISPLAYSECONDARY Controls which axis features are displayed on

the secondary axis.

GRIDATTRS Specifies the attributes of the grid lines.

GRIDDISPLAY Specifies whether axis grid lines are

displayed.

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the

axis label.

LABELFITPOLICY Specifies a policy for fitting axis labels in the

available space.

LABELPOSITION Specifies the position of the axis label.

LABELSPLITCHAR Specifies one or more characters on which the

axis labels can be split, if needed.

LABELSPLITCHARDROP Specifies whether the split characters should

be included in the axis labels that are
displayed.

LABELSPLITJUSTIFY Specifies the justification of the strings that

are inside the axis label blocks.

LINEAROPTS Specifies features for a standard numeric

interval axis.

LOGOPTS Specifies features for a log axis.

NAME Assigns a name to an axis for reference in

other statements.

OFFSETMAX Reserves an area at the maximum end of the

axis. No tick marks are displayed in the
reserved area.
Axis Options for LAYOUT LATTICE 965

Statement Option Description

OFFSETMIN Reserves an area at the minimum end of the

axis. No tick marks are displayed in the
reserved area.

REVERSE Specifies whether the axis origin should be

reversed.

SHORTLABEL Specifies an alternate axis label to use if the

default or specified axis label is too long for
the axis length.

TICKVALUEATTRS Specifies the color and font attributes of the

axis tick values.

TICKVALUEHALIGN Specifies the horizontal alignment for all of

the tick values that are displayed on the Y and
Y2 axes.

TICKVALUEVALIGN Specifies the vertical alignment for all of the

tick values that are displayed on the X and X2
axes.

TIMEOPTS Specifies features for a TIME axis.

TYPE Specifies the type of axis to use.

DISCRETEOPTS=(discrete-axis-options)
specifies one or more options for a discrete axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is space
separated.

Interaction This option is ignored if the axis type is not DISCRETE.

See “Options for Discrete Axes Only” on page 975 for the options that
you can use for discrete-axis-options.

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on the primary axis.
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed
NONE
specifies that no axis features are displayed
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL displays the axis label

LINE displays the axis line
966 Chapter 8 • Axis Options in Layouts

TICKS displays the tick marks

TICKVALUES displays the values that are represented by the major tick
marks

Default STANDARD

Tips The default line attributes for the axis line and axis tick marks are defined
in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall chart
can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

See “Details” on page 1009 for more information about the primary and
secondary axes.

DISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which axis features are displayed on the secondary axis. A secondary axis is
not an independent axis. Rather, it mirrors the primary axis. Thus, for this option to
take effect, all plot statements in the layout must map data to the same primary axis.
For example, a secondary X2 axis can be displayed on top in the layout, provided all
plot statements set XAXIS=X to map data to the primary X axis (bottom). Similarly,
a secondary Y2 axis can be displayed to the right in the layout, provided all plot
statements set YAXIS=Y to map data to the primary Y axis (left).
NONE
specifies that no axis features are displayed
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed on the
secondary axis
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL displays the axis label

LINE displays the axis line
TICKS displays the tick marks
TICKVALUES displays the values that are represented by the major tick
marks

Default NONE

Restriction If some plot statements set XAXIS=X and others set XAXIS=X2, both
the X and X2 axis are primary and a secondary X axis cannot be
Axis Options for LAYOUT LATTICE 967

displayed. In that case, this option is ignored. The same applies for the
Y axes.

Interactions This option is ignored if the COLUMNAXIS statement appears within

a COLUMNAXES block and COLUMN2DATARANGE=UNION or
UNIONALL is in effect.

This option is ignored if the COLUMNAXIS statement appears within

a COLUMN2AXES block and COLUMNDATARANGE=UNION or
UNIONALL is in effect.

Tip Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis
grid lines.

See “Details” on page 1009 for more information about the primary and
secondary axes.

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tip On a log axis, this option affects the appearance of the major grid lines
only. It does not affect the appearance of the minor grid lines. To
control the appearance of the minor grid lines on a log axis, use the
MINORGRIDATTRS= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template to
absolutely control the display of grid lines or to allow interaction with the current
style to decide whether grid lines are displayed. When displayed, the grids appear in
all cells.
AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element in
the current style contains DisplayOpts="ON."
AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in the
current style contains DisplayOpts="OFF."
ON
specifies that grid lines are always displayed. The current style has no override.
OFF
specifies that grid lines are never displayed. The current style has no override.
The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines style
element. Most supplied templates use the default setting AUTO_OFF to indicate a
preference for not displaying grid lines, but allowing the style to override.
968 Chapter 8 • Axis Options in Layouts

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no

ON any value yes

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style has
no preference about grid lines and the graphics template setting for grid
lines is always used.

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The list
form implies that all included string literals or dynamics will be concatenated.

Default The default label is derived from the primary plot in the layout. For
more information, see “When Plots Share Data and a Common Axis”
on page 880.

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the axis label is too long to fit along the axis, then it is truncated by
default.

Tip Use the SHORTLABEL= option to specify an alternate axis label to be

used whenever truncation would normally occur.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display the axis label.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

Axis Options for LAYOUT LATTICE 969

LABELFITPOLICY=AUTO | SPLIT | SPLITALWAYS

specifies a policy for fitting axis labels in the available space.
AUTO
uses the short label, when specified, instead of the original label. If the short label
does not fit, then it is clipped. When no short label is specified, the original label
is clipped.
SPLIT
splits the axis label at a split character, which is specified by the
LABELSPLITCHAR= option, only when necessary in order to make the label fit
the available space. The short label is not used. A split does not occur at a split
character if a split is not needed at that location. If the label does not contain any
of the specified split characters, then it is not split. A label that cannot be split or
that does not fit the available space even after splitting might overlap the
adjoining space.
SPLITALWAYS
always split the axis label at every occurrence of a split character, which is
specified by the LABELSPLITCHAR= option. If the label cannot be split, then it
is clipped.

Default AUTO

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

When the overlay layout is nested in a lattice layout, SPLIT is ignored

and AUTO is used instead.

Note When LABELPOSITION=CENTER, the available area is the full

axis, including the axis offsets. When
LABELPOSITION=DATACENTER, the available area is the tick
display area, excluding the axis offsets.

LABELPOSITION=CENTER | DATACENTER
specifies the position of the axis label.
CENTER
centers each row or column axis label in its axis area. For the external Y and Y2
axes, the label is oriented vertically and is centered in the axis area (including the
offsets) of its row. It is positioned to the left of the tick values for the Y axis or to
the right of the axis values for the Y2 axis. For the external X and X2 axes, the
label is centered in the axis area (including the offsets) of its column. It is
positioned below the tick values for the X axis or above the axis values for the
X2 axis.
DATACENTER
centers each row or column axis label in its axis tick display area. For the
external Y and Y2 axes, each label is oriented vertically and is centered in the
axis tick display area (excluding the offsets) of its row. The labels are positioned
to the left of the tick values for the Y axis or to the right of the axis values for the
Y2 axis. For the external X and X2 axes, each label is centered in the axis tick
display area (excluding the offsets) of its column. The labels are positioned
below the tick values for the X axis or above the axis values for the X2 axis.
The following figure shows the CENTER and DATACENTER positions for external
Y axis labels Open and “Close, and external X axis label Month. An offset is applied
to the maximum end of each axis for demonstration purposes.
970 Chapter 8 • Axis Options in Layouts

Default CENTER

LABELSPLITCHAR="character-list"
specifies one or more characters on which the axis labels can be split, if needed.
When multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in the
axis label. In that case, all of the specified split characters together are treated as a
single split character.
When LABELFITPOLICY=SPLIT, if the axis label does not fit the available space,
then it is split on a specified split character only if a split is needed at that point to
make the label fit. In this case, a split might not occur on every split character. When
LABELFITPOLICY=SPLITALWAYS, the axis label is split unconditionally on
every occurrence of a split character. If the axis label does not contain any of the
specified split characters, the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
labelsplitchar="abc"

The LABELSPLIT=TRUE option must also be specified.

Interactions This option has effect only when LABELPOSITION= is CENTER

or DATACENTER.

The LABELSPLITCHARDROP= option specifies whether the split

characters are included in the displayed data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Axis Options for LAYOUT LATTICE 971

Tip Use the LABELSPLITJUSTIFY= option to specify the justification

of the strings in the axis label block.

LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed axis labels.
TRUE
drops the split characters from the axis label display.
FALSE
includes the split characters in the axis label display. When
LABELSPLIT=TRUE and LABELSPLITCHARDROP=FALSE, each split
character remains as the last character in the current line. The characters that
follow the split character, up to and including the next split character, are then
wrapped to the next line.

Default TRUE. The split characters are dropped from the axis label.

Requirement The LABELSPLIT=TRUE option must also be specified.

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

LABELSPLITJUSTIFY=justification
specifies the justification of the strings that are inside the axis label blocks.
justification
CENTER | LEFT | RIGHT
specifies the justification for the X or X2 axis label.
CENTER | TOP | BOTTOM
specifies the justification for the Y or Y2 axis label.

Default CENTER

Requirement LABELSPLIT=TRUE option must also be specified.

Interaction This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

LINEAROPTS=(linear-axis-options)
specifies one or more options for a numeric interval axis. Options must be enclosed
in parentheses. Each option is specified as a name = value pair and each pair is space
separated.

Interaction This option is ignored if the axis type is not LINEAR.

See “Options for Linear Axes Only” on page 983 for the options that you
can use for linear-axis-options.

LOGOPTS=(log-axis-options)
specifies one or more options for a log axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is space
separated.
972 Chapter 8 • Axis Options in Layouts

Interaction This option is ignored if the axis type is not LOG.

See “Options for Log Axes Only” on page 993 for the options that you
can use for log-axis-options.

NAME="string"
assigns a name to an axis for reference in other statements. Currently, it is used only
in an AXISLEGEND statement.

Interactions This option is ignored unless the axis is discrete. The axis can be
discrete by default, or explicitly set to discrete with a TYPE=
DISCRETE setting.

For this option to take effect, an axis legend must be enabled. To

enable an axis legend, the DISCRETEOPTS= option must set the
TICKVALUEFITPOLICY to either EXTRACT or
EXTRACTALWAYS. In addition, an AXISLEGEND statement must
be specified to generate the axis legend.

OFFSETMAX=AUTO | AUTOCOMPRESS | number

reserves an area at the maximum end of the axis. No tick marks are displayed in the
reserved area.
AUTO
reserves just enough area to fully display markers and other graphical features
near the maximum end of an axis.
AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.
number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886

OFFSETMIN=AUTO | AUTOCOMPRESS | number

reserves an area at the minimum end of the axis. No tick marks are displayed in the
reserved area.
AUTO
reserves just enough area to fully display markers and other graphical features
near the maximum end of an axis.
AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.
number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO
Axis Options for LAYOUT LATTICE 973

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886.

REVERSE=TRUE | FALSE
specifies whether tick values should appear in the reverse order.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

SHORTLABEL="string"
specifies an alternate axis label to display when the default label or the label
specified by the LABEL= option is too long to fit the available space.

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the specified label is itself too long for the axis, it is truncated in the
display.

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick values.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TICKVALUEHALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment for all of the tick values that are displayed on the
Y and Y2 axes.

Defaults RIGHT for a Y axis

LEFT for a Y2 axis

Restriction This option is valid for the Y and Y2 axes only.

974 Chapter 8 • Axis Options in Layouts

TICKVALUEVALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment for all of the tick values that are displayed on the X
and X2 axes.

Defaults TOP for an X axis

BOTTOM for an X2 axis

Restriction This option is valid for the X and X2 axes only.

TIMEOPTS=(time-axis-options)
specifies one or more options for a time axis.

Requirements Columns associated with a time axis must be in SAS time, SAS
date, or SAS datetime units and have an associated SAS time, date,
or datetime format.

Options must be enclosed in parentheses. Each option is specified as

a name = value pair and each pair is space separated.

Interaction This option is ignored if the axis type is not TIME.

See “Options for Time Axes Only” on page 1001 for the options that
you can use for time-axis-options.

TYPE=AUTO | DISCRETE | LINEAR | TIME | LOG

specifies the type of axis to use.
AUTO
requests that the axis type be automatically determined, based on the overlay
contents.
DISCRETE
uses a DISCRETE axis if possible. The data for discrete axes can be character or
numeric. You can add a DISCRETEOPTS= option list to customize this axis
type.
LINEAR
uses a LINEAR axis if possible. You can add a LINEAROPTS= option list to
customize this axis type.
Axis Options for LAYOUT LATTICE 975

TIME
uses a TIME axis if possible. Data for this axis must be SAS time, SAS date, or
SAS datetime values. You can add a TIMEOPTS= option list to customize this
axis type.
LOG
uses a LOG axis if possible. You can add a LOGOPTS= option list to customize
this axis type.

Interaction If a log axis is requested and the axis data contains 0 or negative
values, the axis reverts to a linear axis. This outcome can occur for
the response axis of a bar chart, line chart, needle plot, or waterfall
chart when a baseline intercept of 0 or less is specified. It can also
occur for the response axis of a waterfall chart when an initial bar
value of 0 or less is specified. To get a log response axis in those
cases, set the baseline intercept or initial bar value to a positive
value.

Default AUTO

Interactions If this option is set to anything other than AUTO, then plots within the
layout are dropped from the display if their data types or data ranges
do not match the axis type requirements. For more information, see
“Plot Axis Types Must Agree on Common Axes” on page 883.

After the axis type is determined (whether you set a specific type or
AUTO is in effect), you can use only options that are supported by
that axis type. For example, if TYPE=TIME, then only the general
OVERLAY axis options and those available on TIMEOPTS= are
supported.

Options for Discrete Axes Only

The options that are documented in this section can be used with the DISCRETEOPTS=
axis option. The following table provides a summary of the options.

Discrete Axis Option Description

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the
TICKVALUELIST= option.

TICKTYPE Specifies the position of the axis tick mark.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision on an axis.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the text that is to be displayed for

the tick values that are defined in the
TICKVALUELIST= option.

TICKVALUEROTATION Specifies how the tick values are rotated on

the X and X2 axes.
976 Chapter 8 • Axis Options in Layouts

Discrete Axis Option Description

TICKVALUESPLITCHAR Specifies a list of characters on which the tick

values can be split, if needed.

TICKVALUESPLITCHARDROP Specifies whether the split characters are

included in the displayed tick values.

TICKVALUESPLITJUSTIFY Specifies justification of the strings that are

inside the tick value block.

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string values
that are displayed on the axis in place of the values in the TICKVALUELIST=
option. The strings map one-to-one positionally with the values that are listed in the
TICKVALUELIST= option.

Default Determined by the system or by the TICKVALUELIST= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Tip This option should be used with the TICKVALUELIST= option.

The number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

Example The following example specifies the axis tick values 10, 20, 30, and
40, and the tick display values A, B, C, and D:
tickvaluelist=("10" "20" "30" "40");
tickdisplaylist=("A" "B" "C" "D");

TICKTYPE=MIDPOINT | INBETWEEN
specifies the position of the axis tick marks.

MIDPOINT places the tick marks at the midpoint value location.

INBETWEEN places the tick marks half way between adjacent midpoint
locations.

Default MIDPOINT

Restriction This option applies to discrete axes only.

Note Starting with the second maintenance release of SAS 9.4, when
TICKTYPE=INBETWEEN, the outermost tick marks and grid lines at
each end of the axis are not drawn.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness of a
collision-avoidance policy depends on the number of tick values, their length, and
the length of the axis. Which policies are valid depends on the axis on which this
option is used. For the Y and Y2 axes, the following policies are valid:
Axis Options for LAYOUT LATTICE 977

EXTRACT
displays consecutive integers along the axis instead of the actual tick values in
order to represent those tick values. In most cases, this policy is implemented if
the system estimates that a collision might occur. If no collision occurs, then the
actual tick values are displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User's Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User's Guide.

NONE
makes no attempt to avoid collisions between tick values. Tick values are display
even when they collide.
SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space. A split does not occur at a split character if a split is
not needed at that location. If the value does not contain any of the specified split
characters, then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR=

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that is
specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR=

SPLITALWAYSTHIN
same as SPLITALWAYS, except that thinning is performed when long words do
not fit the available space.
SPLITTHIN
same as SPLIT, except that thinning is performed when long words do not fit the
available space.
THIN
eliminates alternate tick values.
For the X and X2 axes, the following policies are valid:
EXTRACT
display consecutive integers along the axis instead of the actual tick values to
represent those tick values. In most cases, this policy is implemented if the
system estimates that a collision might occur. If no collision occurs, then the
actual tick values are displayed on the axis in the normal manner.
978 Chapter 8 • Axis Options in Layouts

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User's Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User's Guide.

NONE
does not attempt to fit tick values that collide.
ROTATE
rotates the tick values if a collision occurs. The TICKVALUEROTATION=
option specifies whether the values are rotated to a 45-degree diagonal or a 90-
degree vertical position. By default, the values are rotated to a 45-degree
diagonal position.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs. The
TICKVALUEROTATION= option specifies whether the values are rotated to a
45-degree diagonal or a 90-degree vertical position. By default, the values are
rotated to a 45-degree diagonal position.
ROTATEALWAYSDROP
attempts the ROTATEALWAYS policy, and then drops the tick values if
collisions still occur.
ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space. A split does not occur at a split character if a split is
not needed at that location. If the value does not contain any of the specified split
characters, then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR=

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that is
specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR=

STAGGER
alternates the tick values between two rows.
STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.
Axis Options for LAYOUT LATTICE 979

STAGGERTRUNCATE
attempts the STAGGER policy, and then the TRUNCATE policy.
THIN
eliminates alternate tick values.
TRUNCATE
shortens the tick values when they exceed a certain number of characters.
TRUNCATEROTATE
attempts the TRUNCATE policy, and then the ROTATE policy.
TRUNCATESTAGGER
attempts the TRUNCATE policy, and then the STAGGER policy.
TRUNCATETHIN
attempts the TRUNCATE policy, and then the THIN policy.

Defaults ROTATE for the X and X2 axes

THIN for the Y and Y2 axes

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format
specifies how to format the values for major tick marks.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Restrictions This option applies only to discrete axes.

Only character formats are supported.

Interaction This option is ignored when the axis tick values are extracted to an
axis legend. See TICKVALUEFITPOLICY=EXTRACT |
EXTRACTALWAYS on page 976.

Tip Use this option when you want to duplicate tick values on an axis.

TICKVALUELIST=(string-list)
specifies the list of tick values that are to be displayed on the axis.
string-list
a space-separated list of values, enclosed in parentheses. You must enclose each
value in the list in quotation marks.
Only the tick values that are included in the string list are displayed on the axis. The
values are displayed in the order in which they are listed. The data values that are not
in the list are dropped. The list can be a subset of the data values. It can also contain
values that are not included in the actual data. A tick value that is not included in the
data appears on the axis, but no data is represented at its tick mark.

Requirements The list of values must be enclosed in parentheses.

Each value must be enclosed in quotation marks and separated from

adjacent values by a blank space.

Notes If the string list contains duplicate values, then the first occurrence
of the duplicated value in the list is honored and the remaining
instances are ignored.
980 Chapter 8 • Axis Options in Layouts

When the values specified in the list are compared with the actual
data values, leading blanks are honored and trailing blanks are
ignored.

Tips You can use this option to subset the axis values or to display the
values in a specific order.

You can use this option to display values on the axis that are not
contained in the data.

Examples The following example specifies the axis tick values Sedan, Sports,
Wagon, and SUV:
tickvaluelist=("Sedan" "Sports" "Wagon" "SUV")

The following example specifies the axis tick values 10, 20, 30, and
40:
tickvaluelist=("10" "20" "30" "40")

TICKVALUEROTATION=DIAGONAL | VERTICAL
specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL rotates the tick values to a 45-degree diagonal position. The X

labels read left to right in a downward direction. The X2 labels
read left to right in an upward direction.
VERTICAL rotates the labels to a 90-degree vertical position. The labels are
always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXIS statements only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESPLITCHAR="character-list"
specifies a list of characters on which the tick values can be split, if needed. When
multiple characters are specified, each character in the list is treated as a separate
split character unless the specified characters appear consecutively in the tick value.
In that case, all of the specified split characters together are treated as a single split
character.
When a tick value collision is detected, the tick value is split at each occurrence of
any of the characters in the character list or all. If all of the split characters occur
consecutively in the tick value, then they are treated as a single split character. If the
tick value does not contain any of the specified characters, then the value is not split.
"character-list"
one or more characters with no delimiter between each character.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the characters a, b, and c, use the following
option:
Axis Options for LAYOUT LATTICE 981

tickvaluesplitchar="abc"

Interactions This option is ignored unless option TICKVALUEFITPOLICY= is

set to SPLIT, SPLITALWAYS, SPLITTHIN, or
SPLITALWAYSTHIN.

The TICKVALUEFITPOLICY= option sets the policy that is used

to manage the split behavior of the tick values.

The TICKVALUESPLITCHARDROP= option specifies whether

the split characters are displayed or dropped from the display.

Notes When multiple characters are specified, the order of characters in

the list is not significant.

The split characters are case sensitive.

Tips Use the TICKVALUESPLITJUSTIFY= option to specify the

justification of the strings in the tick value block.

For the X and X2 axis tick values, use the TICKVALUEVALIGN=

option to specify the vertical alignment of the tick values.

For the Y and Y2 axis tick values, use the TICKVALUEHALIGN=

option to specify the horizontal alignment of the tick values.

Example The following example specifies a blank space, a comma, and an

underscore as split characters:
tickvaluesplitchar=" ,_"

TICKVALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters should be included in the displayed tick values.
The split characters are specified by the TICKVALUESPLITCHAR= option.
TRUE
drops the split characters from the tick value display. The following figure shows
an example in which TICKVALUESPLITCHARDROP=TRUE and three-word,
asterisk-delimited tick values are split on the asterisk character by using the
SPLITALWAYS policy.

Notice that the asterisk delimiter is not displayed.

FALSE
includes the split characters in the tick value display. The fit policy determines
how the characters are displayed. If the display policy is SPLIT or SPLITTHIN
and TICKVALUESPLITCHARDROP=FALSE, then each tick value is split at a
split character only where a split is necessary in order to make the value fit the
available space. A split might not occur at every split character in the tick value.
At each split point, the split character remains as the last character in the current
line. The characters that follow the split character, up to and including the split
character at the next split point, are then wrapped to the following line. This
process repeats until the entire data tick value is displayed. The following figure
shows an example in which TICKVALUESPLITCHARDROP=FALSE and
982 Chapter 8 • Axis Options in Layouts

three-word, asterisk-delimited tick values are split on the asterisk character by

using the SPLIT policy.

Notice that a split occurs on the first asterisk and not at the second. In this case, a
split is not needed at the second asterisk.
If the fit policy is SPLITALWAYS or SPLITALWAYSTHIN, and
TICKVALUESPLITCHARDROP=FALSE, then each tick value is split at every
instance of a split character in the value regardless of whether a split is actually
needed. Each split character remains as the last character in the current line. The
characters that follow each split character, up to and including the next split
character, are then wrapped to the next line. The following figure shows an
example in which TICKVALUESPLITCHARDROP=FALSE and three-word,
asterisk-delimited tick values are split on the asterisk character by using the
SPLITALWAYS policy.

Notice that a split occurs after each asterisk and each asterisk appears at the end
of the line. In this case, three lines are displayed.

Default TRUE

Interactions The TICKVALUESPLITCHAR= option specifies the split character

or characters.

This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUESPLITJUSTIFY=CENTER | LEFT | RIGHT

specifies justification of the strings that are inside the tick value block. The
justification is relative to an individual tick value’s display area and does not affect
the display of tick values that are not split.
Axis Options for LAYOUT LATTICE 983

Defaults CENTER for an X or X2 axis

RIGHT for a Y axis

LEFT for a Y2 axis

Interaction This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

Options for Linear Axes Only

The options that are documented in this section can be used with the LINEAROPTS=
axis option. The following table provides a summary of the options.

Linear Axis Option Description

INTEGER Specifies that evenly spaced integer values are

used for tick marks.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether the minor tick marks are

displayed on the axis.

THRESHOLDMAX Specifies a bias for including one more tick

mark at the maximum end of the axis.

THRESHOLDMIN Specifies a bias for including one more tick

mark at the minimum end of the axis.

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the
TICKVALUELIST= option.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision on an X or X2 axis. This option is
not available in the ROWAXIS statement.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the order of the tick values for a

linear axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification

can extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on

the X and X2 axes.
984 Chapter 8 • Axis Options in Layouts

Linear Axis Option Description

TICKVALUESEQUENCE Specifies the tick values for a linear axis by

start, end, and increment.

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the
TICKVALUEFORMAT= option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAM or HISTOGRAMPARM statement.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.
Axis Options for LAYOUT LATTICE 985

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
The following figure shows the minor grid lines set to light blue, dotted lines on a linear
axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Defaults Four ticks with five intervals in the first maintenance release of SAS
9.4 and earlier releases.

One tick with two intervals starting with the second maintenance
release of SAS 9.4.
986 Chapter 8 • Axis Options in Layouts

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 885

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30
Axis Options for LAYOUT LATTICE 987

Range 0–1

Restriction This option applies to linear axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 885

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string values
that are displayed on the axis in place of the values in the TICKVALUELIST=
option. The strings map one-to-one positionally with the values that are listed in the
TICKVALUELIST= option.

Default The display of tick values is controlled by the

TICKVALUEFORMAT= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Interaction When this option is specified, the TICKVALUEFORMAT= option

is ignored.

Tip This option should be used with the TICKVALUELIST= option.

The number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness of a
collision-avoidance policy depends on the number of tick values, their length, and
the length of the axis. Which policies are valid depends on the axis on which this
option is used. For the Y and Y2 axes, the following policies are valid:
NONE
makes no attempt to avoid collisions between tick values. Tick values are display
even when they collide.
THIN
eliminates alternate tick values.
For the X and X2 axes, the following policies are valid:
988 Chapter 8 • Axis Options in Layouts

ROTATE
rotates the tick values if a collision occurs. The TICKVALUEROTATION=
option specifies whether the values are rotated to a 45-degree diagonal or a 90-
degree vertical position. By default, the values are rotated to a 45-degree
diagonal position.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs. The
TICKVALUEROTATION= option specifies whether the values are rotated to a
45-degree diagonal or a 90-degree vertical position. By default, the values are
rotated to a 45-degree diagonal position.
ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
STAGGER
alternates the tick values between two rows.
STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.
THIN
eliminates alternate tick values.

Default THIN

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.
(format-options)
specifies one or more formatting options for major tick values. Together, these
options provide parameters for determining an optimal format (w.d, Ew.,
BESTw.) for displaying major tick values.
MAXWIDTH=integer
specifies the maximum width for displayed tick values. Values might be
rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values. Values
might be rounded or converted to E notation to fit into this width.

Default 6

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values. The actual number might vary based on other constraints.

Default 2
Axis Options for LAYOUT LATTICE 989

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it to
reduce the tick value width. The scale can be a named scale or a scientific-
notation scale. The EXTRACTSCALETYPE= option specifies the scale type.
The scale that is used is appended to the axis label, as shown in the following
example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the label
is truncated, and the scale is appended to the truncated label. Ellipses indicate
that the label was truncated, as shown in the following example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis. In
that case, the tick value format might differ from the data
format even when a named format is applied to the data values.

See “boolean ” on page 1339 for other Boolean values that you
can use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.
DEFAULT
extracts a named scale. A named scale can be millions, billions, or
trillions for values of 999 trillion or less, or a multiple of 10 (denoted as
10^n) for values over 999 trillion. For large tick values, the scale factor is
set to ensure that the absolute value of the largest value is greater than 1.
For small fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.
SCIENTIFIC
extracts a scientific-notation scale. A scientific-notation scale is a
multiple of 10 expressed as 10^n for values greater than 1, or a multiple
of 1/10 expressed as 10^–n for values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.
990 Chapter 8 • Axis Options in Layouts

DATA
uses the format that has been assigned to the column that is contributing to the
axis (or BEST6 if no format is assigned) in order to control the formatting of the
major tick values.
format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For details,
see Appendix 4, “SAS Formats Not Supported,” on page 1353.

Note If you specify a format that significantly reduces precision, because

of tick-value rounding, the plot data elements might not align
properly with the axis tick values. In that case, specify a tick-value
format with a higher precision.

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

Interaction This option is ignored when the TICKDISPLAYLIST= option is

specified.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default An internal algorithm determines the tick marks, based on the actual
axis data range or the data range established by the VIEWMIN= and
VIEWMAX= options. By default, when this option is used, the only
tick values that appear are the tick values in numeric-list that fall
within the explicit data range (set by VIEWMIN= and VIEWMAX=)
or the implicit data range (set by the actual data minimum and data
maximum).

Restriction This option applies to linear axes only.

Requirement The tick values must be specified as a space-separated list of numeric

values, enclosed in parentheses.

Interactions This option overrides the INTEGER= option.

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option or the
DISPLAYSECONDARY= option does not display tick values.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then all
ticks in the tick list are displayed. This might result in some data not
being displayed. For example, data might not be displayed when the
VIEWMIN= value is greater than the actual data minimum, or when
the VIEWMAX= value is less than actual data maximum.

If TICKVALUEPRIORITY= TRUE, then the VIEWMIN= and

VIEWMAX= options are ignored if they are fully enclosed by the
numeric-list. The tick numeric-list can extend the implicit data range
of the axis, but cannot reduce it.
Axis Options for LAYOUT LATTICE 991

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option. If the minimum and maximum of the user-
specified values are within the data range, this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Restriction This option applies to linear axes only.

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | VERTICAL
specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL rotates the tick values to a 45-degree diagonal position. The X

labels read left to right in a downward direction. The X2 labels
read left to right in an upward direction.
VERTICAL rotates the labels to a 90-degree vertical position. The labels are
always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXIS statements only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESEQUENCE=(sequence-options)
specifies the tick values by start, end, and increment.
992 Chapter 8 • Axis Options in Layouts

(sequence-options)
a space-separated list of the following name-value-pair options that control major
tick values. You must provide all three options.
START=number
specifies the value for the first tick mark.
END=number
specifies the value for the last tick mark.
INCREMENT=number
specifies the increment for intermediate tick marks between the first and last
tick marks. The END value always controls the last tick mark. The interval
between the last tick mark and the previous tick mark might not necessarily
be the INCREMENT value.

Default An internal algorithm determines the tick marks, based on the actual
axis data range or the data range established by the VIEWMIN= and
VIEWMAX= options. By default, when this option is used, the only
tick values that appear are those that fall within the explicit data range
(set by VIEWMIN= and VIEWMAX=) or the implicit data range (set
by the actual data minimum and data maximum).

Interactions This option overrides the INTEGER= option.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the START= option value and the
VIEWMAX= option is set to the END= option value, then all ticks in
the tick sequence are displayed.

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display tick marks.

Tip The values in the sequence are formatted according to the setting for
the TICKVALUEFORMAT= option.

See TICKVALUELIST= option as an alternative for customizing tick

marks.

VIEWMAX=number
specifies the maximum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The maximum value in the data for the specified axis.

Interactions This option does not determine the maximum axis tick value that is
displayed. The THRESHOLDMAX= value is used to determine the
maximum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
Axis Options for LAYOUT LATTICE 993

factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

VIEWMIN=number
specifies the minimum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The minimum value in the data for the specified axis.

Interactions This option does not determine the minimum axis tick value that is
displayed. The THRESHOLDMIN= value is used to determine the
minimum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Options for Log Axes Only

The options that are described in this section can be used with the LOGOPTS= axis
option. The following table provides a summary of the options.

Log Axis Option Description

BASE Specifies the base of the logarithmic scale for

the axis values.

MINORGRID Specifies whether grid lines are displayed at

the minor tick marks.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether minor ticks are displayed.

TICKINTERVALSTYLE Specifies how to scale and format the values

for major tick marks.
994 Chapter 8 • Axis Options in Layouts

Log Axis Option Description

TICKVALUEFORMAT= Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the tick values for a log axis as a

space-separated list.

TICKVALUEPRIORITY Specifies whether the TICKVALUELIST

specification can extend the axis data range.

VALUESTYPE Specifies the scale that the system uses when

interpreting the TICKVALUELIST=,
VIEWMAX=, and VIEWMIN= option values.

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

BASE=10 | 2 | E
specifies the base of the logarithmic scale for the axis values.

Default 10

Restriction This option applies to log axes only.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.
Axis Options for LAYOUT LATTICE 995

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
The following figure shows the minor grid lines set to light blue, dotted lines on a
base-10 log axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Default Eight ticks with nine intervals (BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT. ).

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.
996 Chapter 8 • Axis Options in Layouts

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKINTERVALSTYLE=AUTO | LOGEXPAND | LOGEXPONENT | LINEAR

specifies how to scale and format the values for major tick marks.
AUTO
selects a LOGEXPAND, LOGEXPONENT, or LINEAR representation
automatically based on the range of the data. When the data range is small
(within an order of magnitude), a LINEAR representation is typically used. Data
ranges that encompass several orders of magnitude typically use the
LOGEXPAND or LOGEXPONENT representation.
LOGEXPAND
places the major tick marks at uniform intervals at integer powers of the base.
The tick values are expanded as follows:

Base=10

Base=2

Base=E

LOGEXPONENT
places the major tick marks at uniform intervals at integer powers of the base.
The tick values are only the integer exponents for all bases.

LINEAR
places the major tick marks at non-uniform intervals that cover the range of the
data.
Axis Options for LAYOUT LATTICE 997

Default AUTO

Restrictions This option applies to log axes only.

For LOGEXPONENT, formats on data columns contributing to the

axis are ignored. For LOGEXPAND, formats on data columns
contributing to the axis are ignored, although any "named format" on
the column is retained. For LINEAR, ticks values are automatically
formatted when the column format is not assigned or one of w.d, Ew.,
or BESTw. Other formats (SAS defined or user-defined) are used if
specified.

GTL currently honors most but not every SAS format. For details, see
Appendix 4, “SAS Formats Not Supported,” on page 1353.

Note When BASE=10 and LOGEXPAND or LOGEXPONENT is used, an

intermediate tick is added whenever the axis data range is less than or
equal to 1.5 powers of 10.

Tip If you use TICKINTERVALSTYLE=LOGEXPONENT, then you

might want to include information in the axis label about which base
is used.

TICKVALUEFORMAT=DATA | format
specifies how to format the values for major tick marks.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
DATA
uses the format that has been assigned to the column that is contributing to the
axis (or BEST6 if no format is assigned) in order to control the formatting of the
major tick values.
format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For details,
see Appendix 4, “SAS Formats Not Supported,” on page 1353.

Restriction This option applies to log axes only.

Interactions This option is ignored when

TICKINTERVALSTYLE=LOGEXPONENT.

When TICKINTERVALSTYLE=LOGEXPAND, this option is

honored for the base 10 and base 2 logarithmic scales, and is ignored
for the base E scale.

When TICKINTERVALSTYLE=LINEAR, this option is honored for

the base 10, base 2, and base E logarithmic scales.

See BASE=

TICKINTERVALSTYLE=
998 Chapter 8 • Axis Options in Layouts

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default Only the tick values specified in the list that fall within the explicit
data range set by the VIEWMIN= and VIEWMAX= options or by
the implicit data range set by the actual data minimum and data
maximum are displayed. An internal algorithm determines the tick
marks.

Requirements The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

The values that you specify must be appropriate for the

VALUESTYPE= specification. Otherwise, unexpected results might
occur. If VALUESTYPE=EXPANDED is in effect (default), specify
increments of the log base power such as 0.1, 1, 10, 100, and so on,
on a base 10 log axis, for example. If VALUESTYPE=EXPONENT
is in effect, specify integer increments of the log base power
exponent such as 1, 2, 3, and so on.

Interactions The VALUESTYPE= option determines how the values in the list
are interpreted.

The VIEWMIN= and VIEWMAX= options alter the axis data

range. If the VIEWMIN= option is set to the minimum tick list
value and the VIEWMAX= option is set to the maximum tick list
value, then all ticks in the tick list are displayed. This might result in
some data not being displayed. For example, data might not be
displayed when the VIEWMIN= value is greater than the actual data
minimum, or when the VIEWMAX= value is less than actual data
maximum.

If the VIEWMIN= value is greater than the actual data minimum or

the VIEWMAX= value is less than actual data maximum, some data
might not be displayed.

This option is ignored if the DISPLAY= or the

DISPLAYSECONDARY= option does not display the tick values.

See TICKINTERVALSTYLE= for specifying the scale and format of the

major tick values

TICKVALUEPRIORITY= for controlling the behavior of the

TICKVALUELIST= option

BASE= for specifying the log base

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by the TICKVALUELIST= option. If the
minimum and maximum of the user-specified values are within the data range,
this option has no effect.
Axis Options for LAYOUT LATTICE 999

FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

VALUESTYPE=EXPANDED | EXPONENT
specifies the scale that the system uses when interpreting the TICKVALUELIST=,
VIEWMAX=, and VIEWMIN= option values. Use this option to choose your
preferred way of specifying log-axis values.

EXPANDED values are interpreted as powers of the base such as 0.1, 1, 10,
100, and so on, for base 10, for example.
EXPONENT values are interpreted as integer exponents of the base such as 1,
2, 3, and so on, for base 10, base 2, and base E.

Default EXPANDED

Note This option does not change the style of the tick values that are
displayed on the axis. It changes only how the VIEWMIN=,
VIEWMAX=, and TICKVALUELIST= option values are interpreted by
the system.

Tip This option is particularly useful when BASE=E.

Examples The following example specifies VIEWMIN= and VIEWMAX= as

exponent values instead of as expanded values on an expanded Base 10
log axis. This results in X-axis tick values of 10, 100, 1000, 10000, and
100000.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexpand
valuestype=exponent
viewmin=1 viewmax=5));

The following example specifies TICKVALUELIST= as a list of

expanded values instead of exponent values on an exponent Base 10 log
axis. This results in X-axis tick values of 1, 2, 3, 4, and 5.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexponent
1000 Chapter 8 • Axis Options in Layouts

tickvaluepriority=true
valuestype=expanded
tickvaluelist=(10 100 1000 10000 100000)));

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Requirement The value that you specify must be appropriate for the
VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is in
effect (default), specify an increment of the log base power such as
0.1, 1, 10, 100, and so on, on a base 10 log axis, for example. If
VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

If an invalid value is specified for the VIEWMAX= option, the

default value for VIEWMAX= is used instead. In that case, if the
default value for VIEWMAX= is less than the value specified by the
VIEWMIN= option, then the VIEWMIN= and VIEWMAX= values
are swapped.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to 1.5
powers of 10.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Examples The following example specifies a value of 100,000 as an expanded

value on a base 10 log axis:
VIEWMAX=100000

The following example specifies a value of 100,000 as an exponent

value on a base 10 log axis:
VIEWMAX=5

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.
Axis Options for LAYOUT LATTICE 1001

Requirement The value that you specify must be appropriate for the
VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is in
effect (default), specify an increment of the log base power such as
0.1, 1, 10, 100, and so on, on a base 10 log axis, for example. If
VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to 1.5
powers of 10.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Examples The following example specifies a value of 0.1 as an expanded value

on a base 10 log axis:
VIEWMIN=0.1

The following example specifies a value of 0.1 as an exponent value

on a base 10 log axis:
VIEWMIN=-1

Options for Time Axes Only

The options that are documented in this section can be used with the TIMEOPTS= axis
option. The following table provides a summary of the options.

Time Axis Option Description

INTERVAL Specifies the time interval between major tick

marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time

interval that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

1002 Chapter 8 • Axis Options in Layouts

Time Axis Option Description

MINORTICKINTERVAL Specifies the time interval between minor

ticks.

MINORTICKS Specifies whether minor tick marks are

displayed.

SPLITTICKVALUE Specifies whether to split the tick values on

column axes, if possible. This option is not
available in the ROWAXIS statement.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision on column axes. This option is not
available in the ROWAXIS statement.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the order of the tick values for a

time axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification

can extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on

the X and X2 axes.

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

INTERVAL=interval
specifies the time interval between major ticks. Valid interval keywords are AUTO,
SECOND, MINUTE, HOUR, DAY, TENDAY, WEEK, SEMIMONTH, MONTH,
QUARTER, SEMIYEAR, YEAR.
Table 8.3 Time Intervals

Default tick value

INTERVAL Unit Tick interval format

AUTO DATE, TIME, or automatically automatically

DATETIME chosen chosen

SECOND TIME or second TIME8.

DATETIME

MINUTE TIME or minute TIME8.

DATETIME

HOUR TIME or hour TIME8.

DATETIME
Axis Options for LAYOUT LATTICE 1003

Default tick value

INTERVAL Unit Tick interval format

DAY DATE or day DATE9.

DATETIME

TENDAY DATE or 10 days DATE9.

DATETIME

WEEK DATE or 7 days DATE9.

DATETIME

SEMIMONTH DATE or 1st and 16th of each DATE9.

DATETIME month

MONTH DATE or month MONYY7.

DATETIME

QUARTER DATE or 3 months YYQC6.

DATETIME

SEMIYEAR DATE or 6 months MONYY7.

DATETIME

YEAR DATE or year YEAR4.

DATETIME

Default AUTO. An appropriate interval is chosen based on the data and the
column date, date-time, or time format.

Restriction This option applies to time axes only.

Requirement The data column(s) mapped to a time axis must be in the same
duration units: TIME, DATE, or DATETIME. The selection of an
interval must be consistent with the duration unit. For example, if the
data are in time units, you can specify only AUTO, SECOND,
MINUTE, HOUR.

Interaction This option is ignored if the TICKVALUELIST= option is used.

INTERVALMULTIPLIER=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the axis.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default 1

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= option is used.

Tip Use the INTERVAL= option to specify a different time interval.

Examples To specify 3-month intervals:

1004 Chapter 8 • Axis Options in Layouts

INTERVAL=MONTH INTERVALMULTIPLIER=3

To specify 10-year intervals:

INTERVAL=YEAR INTERVALMULTIPLIER=10

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
The following figure shows the minor grid lines set to light blue, dotted lines on a time
axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.
Axis Options for LAYOUT LATTICE 1005

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKINTERVAL=interval
specifies the time interval between minor ticks. See Table 8.3 on page 1002 for
information about the intervals that you can select. The interval that you select must
be consistent with the axis data duration units such as TIME, DATE, or DATETIME.
For example, if the axis data is in TIME units, then you must specify AUTO,
SECOND, MINUTE, or HOUR.

Default AUTO

Interactions This option is ignored if the TICKVALUELIST= option is used.

This option is ignored if the MINORTICKINTERVAL= setting is

greater than the INTERVAL= setting.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Interactions The number of minor ticks is dependent on the value of the

MINORTICKINTERVAL= option, if specified. If
MINORTICKINTERVAL= is not specified, then it is dependent on the
value of the INTERVAL= option.

This option is ignored if the TICKVALUELIST= option is used, or if

the DISPLAY= or DISPLAYSECONDARY= option does not display
the tick marks.
1006 Chapter 8 • Axis Options in Layouts

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

SPLITTICKVALUE=TRUE | FALSE
specifies whether to split the tick values on an X or X2 axis, if possible. This option
is not available for a Y or Y2 axis.
TRUE
splits the axis tick values into two lines allowing more tick values to appear. For
example, with INTERVAL= MONTH, this is how tick values are split:

FALSE
does not split the axis tick values. For example, when this option specifies
FALSE, this is how the tick values in the previous example appear:

Typically, fewer tick values fit, causing thinning, rotation, or staggering of the
values. See the TICKVALUEFITPOLICY= option.

Default TRUE

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUEFORMAT= option is used.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an X or X2 axis. This option is
not available for the Y and Y2 axes. The effectiveness of a collision-avoidance
policy depends on the number of tick values, their length, and the length of the axis.
The following policies are valid:
THIN
eliminates alternate tick values.
ROTATE
rotates the tick values if a collision occurs. The TICKVALUEROTATION=
option specifies whether the values are rotated to a 45-degree diagonal or a 90-
degree vertical position. By default, the values are rotated to a 45-degree
diagonal position.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs. The
TICKVALUEROTATION= option specifies whether the values are rotated to a
45-degree diagonal or a 90-degree vertical position. By default, the values are
rotated to a 45-degree diagonal position.
Axis Options for LAYOUT LATTICE 1007

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
STAGGER
alternates the tick values between two rows.
STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

Default THIN

Restriction This option is valid only for the X and X2 axes.

Interaction When SPLITTICKVALUE= TRUE, this option is ignored and only the
THIN policy is used.

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format | DATA
specifies how to format the values for major tick marks.
format
specifies a SAS date, time, or datetime format to control how the major tick
values are displayed. This format must be in the same duration units as the data
column(s) mapped to a time axis: TIME, DATE, or DATETIME and should be
appropriate for the value of the INTERVAL= option. For example, if
INTERVAL=MONTH and there are two years of data displayed on the axis, then
choosing TICKVALUEFORMAT=YEAR. would result in several ticks having
the same year value.
DATA
specifies that the SAS date, time, or datetime format associated with the data
column assigned to the axis be used to control how the major tick values are
displayed.

Default The default format used by the INTERVAL= option. The default does
not apply if TICKVALUELIST= is specified.

Restrictions This option applies to time axes only.

GTL currently honors most but not every SAS format. For details, see
Appendix 4, “SAS Formats Not Supported,” on page 1353.

TICKVALUELIST=(time-constant-list | date-constant-list | datetime-constant-list |

numeric-list)
specifies the tick values for a time axis as list.

Default An internal algorithm determines the tick values.

Restrictions This option applies to time axes only.

If TICKVALUEPRIORITY= is set to FALSE, then this option does

not extend the data range of the axis. If the values fall within the
default data range or that specified by the VIEWMIN= or
VIEWMAX= options, then they are used.
1008 Chapter 8 • Axis Options in Layouts

Requirement The tick values must be specified as a space-separated list of values

enclosed in parentheses. The items in the list must be in the same
duration units as the data mapped to the axis: TIME, DATE, or
DATETIME. The values can be expressed as SAS TIME, DATE, or
DATETIME constants (for example, "13:23"T, "11MAY06"D, or
"11MAY06:13:23"DT) or their numeric equivalents.

Interactions The values in the list are formatted according to the format specified
on the TICKVALUEFORMAT= option. If TICKVALUEFORMAT=
is not used, then the values are formatted according to the column
format (the default TICKVALUEFORMAT value is not applied to
these values).

If this option is specified, the SPLITTICKVALUE= and

INTERVAL= options are ignored.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by the TICKVALUELIST= option. If the
minimum and maximum of the user-specified values are within the data range,
this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | VERTICAL
specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL rotates the tick values to a 45-degree diagonal position. The X

labels read left to right in a downward direction. The X2 labels
read left to right in an upward direction.
VERTICAL rotates the labels to a 90-degree vertical position. The labels are
always drawn from bottom to top.
Axis Options for LAYOUT LATTICE 1009

Default DIAGONAL

Restriction This option is valid for COLUMNAXIS statements only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the original
data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX= value.
The VIEWMIN= and VIEWMAX= values, and additional factors such as
thresholds and the tick values computed by the plot statement, are used to
determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the original
data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN= value. The
VIEWMIN= and VIEWMAX= values, and additional factors such as
thresholds and the tick values computed by the plot statement, are used to
determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Details
The LAYOUT LATTICE statement creates a grid of graphs that automatically aligns plot
areas, data display areas, axis labels, and headers across the columns and rows of the
layout. The axis data ranges can be scaled, and the axes for individual cells in the layout
can be managed by row and by column using COLUMNAXIS and ROWAXIS
statements. COLUMNAXIS statements are used within a COLUMNAXES or
COLUMN2AXES block to externalize column axes for the layout. Similarly,
ROWAXIS statements are used within a ROWAXES or ROW2AXES block to
externalize row axes for the layout. Each axis block is used to manage the primary axis.
The axis that is considered primary depends on the settings for the XAXIS= and
YAXIS= options in plot statements that are specified within the layout:
1010 Chapter 8 • Axis Options in Layouts

Option Setting Primary Axis Axis Block to Use

XAXIS=X X (bottom) COLUMNAXES

XAXIS=X2 X2 (top) COLUMN2AXES

YAXIS=Y Y (left) ROWAXES

YAXIS=Y2 Y2 (right) ROW2AXES

For the specifications to take effect,

• UNION or UNIONALL data scaling must be set for the affected columns and rows.
The data scaling is set with the LAYOUT LATTICE statement’s
COLUMNDATARANGE=, COLUMN2DATARANGE=, ROWDATARANGE=,
and ROW2DATARANGE= options.
• Within a COLUMNAXES or COLUMN2AXES block, one COLUMNAXIS
statement should be specified for each column that contains axes that you need to
manage. Both axes blocks can contain a COLUMNAXIS statement for the same
column. For example, to manage the axes in the first column of the layout, the
COLUMNAXES block can contain a COLUMNAXIS statement that manages the
column’s X axes. The COLUMN2AXES block can contain a COLUMNAXIS
statement that manages the column’s X2 axes.
• Within a ROWAXES or ROW2AXES block, one ROWAXIS statement should be
specified for each row that contains axes that you need to manage. Both axes blocks
can contain a ROWAXIS statement for the same row. For example, to manage the
axes in the first row of the layout, the ROWAXES block can contain a ROWAXIS
statement that manages the row’s Y axes. The ROW2AXES block can contain a
ROWAXIS statement that manages the column’s Y2 axes.
In addition to managing the primary axes, you can also display “secondary” axes in the
grid. A secondary axis is not an independent axis. Rather, it mirrors the primary axis, but
it is displayed on the opposite side and can have different display options. For example,
when the X axis (bottom) is primary, you can mirror that axis with a secondary X axis at
the top of the grid. Similarly, when the Y2 axis (right) is primary, you can mirror that
axis with a secondary Y2 axis on the left of the grid. A secondary axis makes it easier to
interpolate values in the cells that are farthest away from the primary axis. To display a
secondary axis, use the DISPLAYSECONDARY= option.
For general information about managing primary and secondary axes, see “Plot Data Are
Mapped to a Designated Axis” on page 876. For details about managing the axes within
a LAYOUT LATTICE, see the discussion about the LAYOUT LATTICE “Axis
Statements” on page 129.
The following example shows COLUMNAXIS statements for a lattice with two
columns:
layout lattice / columns=2 columndatarange=union;
columnaxes;
columnaxis / griddisplay=on displaysecondary=(ticks tickvalues);
columnaxis / griddisplay=on displaysecondary=(ticks tickvalues);
endcolumnaxes;

/* rest of lattice definition */

endlayout;
Example: Axis Options for LAYOUT LATTICE 1011

COLUMNAXIS and ROWAXIS statements are similar to the XAXISOPTS= and

YAXISOPTS= options for LAYOUT OVERLAY, with the following differences:
• When COLUMNAXIS and ROWAXIS are used, any axis options specified on plots
within the affected columns or rows are ignored. All axis features for the external
axes must be specified on the COLUMNAXIS or ROWAXIS statement.
• When COLUMNAXIS and ROWAXIS are used, any LAYOUT
OVERLAYEQUATED layouts specified for cells in the affected columns or rows are
implemented as LAYOUT OVERLAY layouts. Equated axes are not supported on
external axes.
In the default cases for each plot in the layout, the axis type is always DISCRETE,
LINEAR, or TIME. The TYPE= option enables you to specify an axis type that
overrides the default. For example, when appropriate for the data, you can request a
LOG axis. When you override the default axis type, you must be sure to specify the
correct axis type for the plot(s) that you are defining.
Each axis type has features specific to that type, and the following axis options enable
you to specify features for the different types: DISCRETEOPTS= , LINEAROPTS= ,
LOGOPTS= , and TIMEOPTS= . One or more of these options can be specified for an
axis, but the specified settings are applied only to the axis type that supports them.

Example: Axis Options for LAYOUT LATTICE

This example shows how to externalize axes in a LAYOUT LATTICE and manage the
axis features on primary Y and Y2 axes.
• The first HISTOGRAM statement specifies YAXIS=Y2 to make the Y2 axis the
primary axis for COUNT measures. The second HISTOGRAM statement specifies
YAXIS=Y to make the Y axis the primary axis for PERCENT measures.
• In order to externalize the axes within the layout, the data ranges for the axes must be
unified. In the LAYOUT LATTICE statement, the ROWDATARANGE= option
unifies the data ranges for the Y axes across the row. The ROW2DATARANGE=
option unifies the data ranges for the Y2 axes across the row.
• The ROWAXIS statement is used to manage axis features for the row axes. To
manage the primary Y axis, a ROWAXIS statement is specified within a ROWAXES
block. To manage the primary Y2 axis, another ROWAXIS statement is specified
within a ROW2AXES block.
• Within the ROWAXES block, the ROWAXIS statement consolidates Y axes in the
row into a single, external Y axis and also displays grid lines. Within the
ROW2AXES block, the ROWAXIS statement consolidates Y2 axes in the row into a
single, external Y2 axis, but it does not alter the default features of that axis.
The following graph was generated by the “Example Program” on page 1012 :
1012 Chapter 8 • Axis Options in Layouts

Example Program
proc template;
define statgraph y2axis;
begingraph;
layout lattice / columns=2 columngutter=10
rowdatarange=union row2datarange=union ;
rowaxes;
rowaxis / griddisplay=on;
endrowaxes;
row2axes;
rowaxis;
endrow2axes;
layout overlay;
histogram height / scale=count yaxis=y2 ;
histogram height / scale=percent yaxis=y ;
densityplot height / normal();
endlayout;
layout overlay;
histogram weight / scale=count yaxis=y2 ;
histogram weight / scale=percent yaxis=y ;
densityplot weight / normal();
endlayout;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=y2axis;

run;
Axis Options for LAYOUT OVERLAYEQUATED 1013

Axis Options for LAYOUT OVERLAYEQUATED

Axis options for the plots within an OVERLAYEQUATED layout.
Interaction: The OVERLAYEQUATED’s axis options are ignored when the LAYOUT
OVERLAYEQUATED statement is nested within another layout type that has
external axes in effect. For example, the axis options are ignored when the
statement is nested within a LAYOUT LATTICE with a COLUMNAXIS= or
ROWAXIS= option in effect.
See: “LAYOUT OVERLAYEQUATED Statement” on page 144

Syntax
Axis options for the plots within an OVERLAYEQUATED layout are specified with the
following options on a LAYOUT OVERLAYEQUATED statement:
COMMONAXISOPTS=(common-equated-axis-options)
XAXISOPTS=(equated-axis-options)
YAXISOPTS=(equated-axis-options)

Options That Apply in Common to Both Equated Axes

The options that are documented in this section are specified with the
COMMONAXISOPTS= option and are applied to both the X and Y axes. With the
exception of VIEWMAX and VIEWMIN, these options cannot be applied separately to
an X or Y axis using the XAXISOPTS= or YAXISOPTS= option. See “Options That
Apply Separately to an X or Y Equated Axis” for a list of options that can be applied to a
single axis. The following table provides a summary of the common options.

Equated Axis Option Description

INTEGER Specifies that evenly spaced integer values are

used for tick marks for all axes.

LINEEXTENT Specifies the extent of the axis lines.

TICKSTYLE Specifies the placement of tick marks in

relation to the axis line.

TICKVALUELIST Specifies the order of the tick values as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification

can extend the axis data range.

TICKVALUESEQUENCE Specifies the tick values by start, end, and

increment.

VIEWMAX Specifies the maximum data value to include

in the display on the X and Y axes when the
axis lengths and major tick values are equal.
1014 Chapter 8 • Axis Options in Layouts

Equated Axis Option Description

VIEWMIN Specifies the minimum data value to include

in the display on the X and Y axes when the
axis lengths and major tick values are equal.

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the
TICKVALUEFORMAT= option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAM or HISTOGRAMPARM statement.

See “boolean ” on page 1339 for other Boolean values that you can use.

LINEEXTENT=FULL | DATA | number

specifies the extent of the axis lines.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
FULL
specifies axis lines that extend along the entire length of the axis.
DATA
specifies axis lines that extend through the data range from the minimum offset
to the maximum offset.
number
specifies how much the axis lines extend from DATA toward FULL as a decimal
proportion. A value of 0 is equivalent to DATA, and a value of 1 is equivalent to
FULL.

Range 0–1

Tip A numeric value is useful for bar charts when DATA terminates the axis
line at the midpoint positions of the minimum and maximum bars. In
that case, you can specify a numeric value to lengthen the axis line so
that it extends to the full width of both bars.
Axis Options for LAYOUT OVERLAYEQUATED 1015

The following figure shows a simple example of each value for the X and Y axis lines.
The light-blue dashed lines depict the minimum and maximum offsets that are set on the
axes.

Default FULL

Restriction This option is valid only in OVERLAY and OVERLAYEQUATED

layouts.

Interaction This option overrides the AXISLINEEXTENT= option in the

BEGINGRAPH statement.

Tip The graph wall outline might appear to be an axis line. In that case, use
the WALLDISPLAY=NONE or WALLDISPLAY=(FILL) option in the
layout statement to suppress the wall outline.

TICKSTYLE=OUTSIDE | INSIDE | ACROSS

specifies the placement of tick marks in relation to the axis line. The figure shows
the tick display for each value.

OUTSIDE displays tick marks outside of the axis frame.

INSIDE displays tick marks inside the axis frame.
ACROSS displays tick marks across the axis line.

Default The GraphAxisLines:TickDisplay style reference.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick marks.

Notes This option has no affect on the placement of the tick values, which are
always outside the axis frame.

This option applies to both major ticks and minor ticks.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.
1016 Chapter 8 • Axis Options in Layouts

Default An internal algorithm determines the tick marks, based on the actual
axis data range or the data range established by the VIEWMIN= and
VIEWMAX= options. By default, when this option is used, the only
tick values that appear are the tick values in numeric-list that fall
within the explicit data range (set by VIEWMIN= and VIEWMAX=)
or the implicit data range (set by the actual data minimum and data
maximum).

Requirement The tick values must be specified as a space-separated list of numeric

values, enclosed in parentheses.

Interactions This option overrides the INTEGER= option.

This option is ignored if the LAYOUT OVERLAYEQUATED

statement specifies EQUATETYPE=FIT (the default).

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option or the
DISPLAYSECONDARY= option does not display tick values.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then all
ticks in the tick list are displayed. This might result in some data not
being displayed. For example, data might not be displayed when the
VIEWMIN= value is greater than the actual data minimum, or when
the VIEWMAX= value is less than actual data maximum.

If TICKVALUEPRIORITY= TRUE, then the VIEWMIN= and

VIEWMAX= options are ignored if they are fully enclosed by the
numeric-list. The tick numeric-list can extend the implicit data range
of the axis, but cannot reduce it.

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option. If the minimum and maximum of the user-
specified values are within the data range, this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.
Axis Options for LAYOUT OVERLAYEQUATED 1017

This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUESEQUENCE=(sequence-options)
specifies the tick values by start, end, and increment.
(sequence-options)
a space-separated list of the following name-value-pair options that control major
tick values. You must provide all three options.
START=number
specifies the value for the first tick mark.
END=number
specifies the value for the last tick mark.
INCREMENT=number
specifies the increment for intermediate tick marks between the first and last
tick marks. The END value always controls the last tick mark. The interval
between the last tick mark and the previous tick mark might not necessarily
be the INCREMENT value.

Default An internal algorithm determines the tick marks, based on the actual
axis data range or the data range established by the VIEWMIN= and
VIEWMAX= options. By default, when this option is used, the only
tick values that appear are those that fall within the explicit data range
(set by VIEWMIN= and VIEWMAX=) or the implicit data range (set
by the actual data minimum and data maximum).

Interactions This option overrides the INTEGER= option.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the START= option value and the
VIEWMAX= option is set to the END= option value, then all ticks in
the tick sequence are displayed.

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the LAYOUT OVERLAYEQUATED

statement specifies EQUATETYPE=FIT (the default), or if the
DISPLAY= option or the DISPLAYSECONDARY= option does not
display tick marks.

Tip The values in the sequence are formatted according to the setting for
the TICKVALUEFORMAT= option.

See TICKVALUELIST= option as an alternative for customizing tick

marks.
1018 Chapter 8 • Axis Options in Layouts

VIEWMAX=number
specifies the maximum data value to include in the display on the X and Y axes
when the axis lengths and major tick values are equal. The value might be adjusted
by the threshold calculation.

Default The maximum value in the data for the X and Y axes.

Restriction This option is honored only when EQUATETYPE=SQUARE.

Interactions This option does not determine the maximum axis tick value that is
displayed. The THRESHOLDMAX= value is used to determine the
maximum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

VIEWMIN=number
specifies the minimum data value to include in the display on the X and Y axes when
the axis lengths and major tick values are equal. The value might be adjusted by the
threshold calculation.

Default The minimum value in the data for the X and Y axes.

Restriction This option is honored only when EQUATETYPE=SQUARE.

Interactions This option does not determine the minimum axis tick value that is
displayed. The THRESHOLDMIN= value is used to determine the
minimum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Axis Options for LAYOUT OVERLAYEQUATED 1019

Options That Apply Separately to an X or Y Equated Axis

The options that are documented in this section can be applied to an X axis with the
XAXISOPTS= option, or to the Y axis with the YAXISOPTS= option. See “Options
That Apply in Common to Both Equated Axes” on page 1013 for a list of options that
apply in common to both axes. The following table provides a summary of the options.

Equated Axis Option Description

DISPLAY Controls which axis features are displayed on

the primary axis.

DISPLAYSECONDARY Controls which axis features are displayed on

the secondary axis.

GRIDATTRS Specifies the attributes of the grid lines.

GRIDDISPLAY Specifies when axis grid lines are displayed.

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the

axis label.

LINEEXTENT Specifies the extent of the axis line.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether the minor tick marks are

displayed on the axis.

OFFSETMAX Reserves an area at the maximum end of the

axis. No tick marks are displayed in the
reserved area.

OFFSETMIN Reserves an area at the minimum end of the

axis. No tick marks are displayed in the
reserved area.

REVERSE Specifies whether tick values should appear in

the reverse order.

SHORTLABEL Specifies an alternate axis label to use if the

default or specified axis label is too long for
the axis length.

THRESHOLDMAX Specifies a bias for including one more tick

mark at the maximum end of the axis.
1020 Chapter 8 • Axis Options in Layouts

Equated Axis Option Description

THRESHOLDMIN Specifies a bias for including one more tick

mark at the minimum end of the axis.

TICKVALUEATTRS Specifies the color and font attributes of the

axis tick values.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision. Only the default policy (THIN) is
available for a Y or Y2 axis.

TICKVALUEFORMAT Specifies how to format the values for tick

marks.

VIEWMAX Specifies the maximum data value to include

in the display on the X or Y axis when the
axis lengths and major tick values are not
equal.

VIEWMIN Specifies the minimum data value to include

in the display on the X or Y axis when the
axis lengths and major tick values are not
equal.

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on the primary axis.
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed
NONE
specifies that no axis features are displayed
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL displays the axis label

LINE displays the axis line
TICKS displays the tick marks
TICKVALUES displays the values that are represented by the major tick
marks

Default STANDARD

Tips The default line attributes for the axis line and axis tick marks are defined
in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.
Axis Options for LAYOUT OVERLAYEQUATED 1021

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall chart
can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

DISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which axis features are displayed on the secondary axis. When data are
mapped to the X or Y axis, you can display an X2 or Y2 (secondary) axis using this
option. The secondary axis is a duplicate of the X or Y axis but can have different
display options.
NONE
specifies that no axis features are displayed
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed on the
secondary axis
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL displays the axis label

LINE displays the axis line
TICKS displays the tick marks
TICKVALUES displays the values that are represented by the major tick
marks

Default NONE

Tip Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template to
absolutely control the display of grid lines or to allow interaction with the current
style to decide whether grid lines are displayed.
1022 Chapter 8 • Axis Options in Layouts

AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element in
the current style contains DisplayOpts="ON."
AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in the
current style contains DisplayOpts="OFF."
ON
specifies that grid lines are always displayed. The current style has no override.
OFF
specifies that grid lines are never displayed. The current style has no override.
The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines style
element. Most supplied templates use the default setting AUTO_OFF to indicate a
preference for not displaying grid lines, but allowing the style to override.

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no

ON any value yes

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style has
no preference about grid lines and the graphics template setting for grid
lines is always used.

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The list
form implies that all included string literals or dynamics will be concatenated.

Default The default label is derived from the primary plot in the layout. For more
information, see “When Plots Share Data and a Common Axis” on page
880.

Note If the axis label is too long to fit along the axis, then it is truncated by
default.
Axis Options for LAYOUT OVERLAYEQUATED 1023

Tip Use the SHORTLABEL= option to specify an alternate axis label to be

used whenever truncation would normally occur.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display the axis label.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LINEEXTENT=FULL | DATA | number

specifies the extent of the axis line.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
FULL
specifies an axis line that extends along the entire length of the axis.
DATA
specifies an axis line that extends through the data range from the minimum
offset to the maximum offset.
number
specifies how much the axis line extends from DATA toward FULL as a decimal
proportion. A value of 0 is equivalent to DATA, and a value of 1 is equivalent to
FULL.

Range 0–1

Tip A numeric value is useful for bar charts when DATA terminates the axis
line at the midpoint positions of the minimum and maximum bars. In
that case, you can specify a numeric value to lengthen the axis line so
that it extends to the full width of both bars.

The following figure shows a simple example of each value for the X and Y axis lines.
The light-blue dashed lines depict the minimum and maximum offsets that are set on the
axes.

Default FULL
1024 Chapter 8 • Axis Options in Layouts

Restriction This option is valid only in OVERLAY and OVERLAYEQUATED

layouts.

Interactions This option overrides the AXISLINEEXTENT= option in the

BEGINGRAPH statement.

This option overrides the LINEEXTENT= option in

COMMONAXISOPTS=.

Tip The graph wall outline might appear to be an axis line. In that case,
use the WALLDISPLAY=NONE or WALLDISPLAY=(FILL) option
in the layout statement to suppress the wall outline.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
Axis Options for LAYOUT OVERLAYEQUATED 1025

The following figure shows the minor grid lines set to light blue, dotted lines on a linear
axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE
1026 Chapter 8 • Axis Options in Layouts

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

OFFSETMAX=AUTO | AUTOCOMPRESS | number

reserves an area at the maximum end of the axis. No tick marks are displayed in the
reserved area.
AUTO
reserves just enough area to fully display markers and other graphical features
near the maximum end of an axis.
AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.
number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886

OFFSETMIN=AUTO | AUTOCOMPRESS | number

reserves an area at the minimum end of the axis. No tick marks are displayed in the
reserved area.
AUTO
reserves just enough area to fully display markers and other graphical features
near the maximum end of an axis.
AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.
number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886.

REVERSE=TRUE | FALSE
specifies whether tick values should appear in the reverse order.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

SHORTLABEL="string"
specifies an alternate axis label to display when the default label or the label
specified by the LABEL= option is too long to fit the available space.
Axis Options for LAYOUT OVERLAYEQUATED 1027

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the specified label is itself too long for the axis, it is truncated in the
display.

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to the LAYOUT OVERLAYEQUATED

XAXISOPTS= and YAXISOPTS= options only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 885

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to the LAYOUT OVERLAYEQUATED

XAXISOPTS= and YAXISOPTS= options only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.
1028 Chapter 8 • Axis Options in Layouts

See “Adjusting Axis Thresholds” on page 885

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick values.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness of a
collision-avoidance policy depends on the number of tick values, their length, and
the length of the axis. Which policies are valid depends on the axis on which this
option is used. For the Y and Y2 axes, the following policies are valid:
NONE
makes no attempt to avoid collisions between tick values. Tick values are display
even when they collide.
THIN
eliminates alternate tick values.
For the X and X2 axes, the following policies are valid:
ROTATE
rotates the tick values if a collision occurs.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.
ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
STAGGER
alternates the tick values between two rows.
STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.
THIN
eliminates alternate tick values.

Default THIN

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.
(format-options)
specifies one or more formatting options for major tick values. Together, these
options provide parameters for determining an optimal format (w.d, Ew.,
BESTw.) for displaying major tick values.
Axis Options for LAYOUT OVERLAYEQUATED 1029

MAXWIDTH=integer
specifies the maximum width for displayed tick values. Values might be
rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values. Values
might be rounded or converted to E notation to fit into this width.

Default 6

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values. The actual number might vary based on other constraints.

Default 2

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it to
reduce the tick value width. The scale can be a named scale or a scientific-
notation scale. The EXTRACTSCALETYPE= option specifies the scale type.
The scale that is used is appended to the axis label, as shown in the following
example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the label
is truncated, and the scale is appended to the truncated label. Ellipses indicate
that the label was truncated, as shown in the following example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis. In
that case, the tick value format might differ from the data
format even when a named format is applied to the data values.

See “boolean ” on page 1339 for other Boolean values that you
can use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.
1030 Chapter 8 • Axis Options in Layouts

DEFAULT
extracts a named scale. A named scale can be millions, billions, or
trillions for values of 999 trillion or less, or a multiple of 10 (denoted as
10^n) for values over 999 trillion. For large tick values, the scale factor is
set to ensure that the absolute value of the largest value is greater than 1.
For small fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.
SCIENTIFIC
extracts a scientific-notation scale. A scientific-notation scale is a
multiple of 10 expressed as 10^n for values greater than 1, or a multiple
of 1/10 expressed as 10^–n for values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.

DATA
uses the format that has been assigned to the column that is contributing to the
axis (or BEST6 if no format is assigned) in order to control the formatting of the
major tick values.
format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For details,
see Appendix 4, “SAS Formats Not Supported,” on page 1353.

Note If you specify a format that significantly reduces precision, because

of tick-value rounding, the plot data elements might not align
properly with the axis tick values. In that case, specify a tick-value
format with a higher precision.

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

VIEWMAX=number
specifies the maximum data value to include in the display on the X or Y axes when
the axis lengths and major tick values are equal. The value might be adjusted by the
threshold calculation.

Default The maximum value in the data for the X and Y axes.

Interactions This option is ignored when EQUATETYPE=SQUARE.

This option does not determine the maximum axis tick value that is
displayed. The THRESHOLDMAX= value is used to determine the
maximum tick value.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
Axis Options for LAYOUT OVERLAYEQUATED 1031

factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

VIEWMIN=number
specifies the minimum data value to include in the display on the X or Y axes when
the axis lengths and major tick values are equal. The value might be adjusted by the
threshold calculation.

Default The minimum value in the data for the X and Y axes.

Interactions This option is ignored when EQUATETYPE=SQUARE.

This option does not determine the minimum axis tick value that is
displayed. The THRESHOLDMIN= value is used to determine the
minimum tick value.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Details
The LAYOUT OVERLAYEQUATED statement is used for equated layouts where the X
and Y axes always have equal increments between tick values. Because the axes within
the equated layout are so closely correlated, some axis adjustments cannot be applied to
one axis without applying them to the other.
For example, the INTEGER= option sets evenly spaced integer values for the axis tick
marks. That setting must be applied to both axes if the correlation between them is to be
maintained. The LAYOUT OVERLAYEQUATED statement provides the
COMMONAXISOPTS= option for specifying the INTEGER= option and other options
whose settings apply in common to both axes.
Despite the close correlation between the axes, some axis adjustments can be made to
one axis without affecting the other. For example, displaying grid lines on one axis has
no impact on the other. The XAXISOPTS= and YAXISOPTS= options are available for
applying settings separately to the X and Y axes.
The following example template sets evenly spaced integer values for the axis tick marks
of both axes. It also specifies the display of grid lines, tick marks, and tick values for the
Y axis:
begingraph;
layout overlayequated /
commonaxisopts=(integer=true);
1032 Chapter 8 • Axis Options in Layouts

yaxisopts=(griddisplay=on display=(ticks tickvalues));

seriesplot x=var1 y=var2;
endlayout;
endgraph;

Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL

Axis options for the plots within DATALATTICE and DATAPANEL layouts
See: “LAYOUT DATALATTICE Statement” on page 45
“LAYOUT DATAPANEL Statement” on page 70

Syntax
The X and X2 axis options for the plots within DATALATTICE and DATAPANEL
layouts are specified with the following options:
COLUMNAXISOPTS=(axis-option(s))
COLUMN2AXISOPTS=(axis-option(s))
The Y and Y2 axis options for the plots within DATALATTICE and DATAPANEL
layouts are specified with the following options:
ROWAXISOPTS=(axis-option(s))
ROW2AXISOPTS=(axis-option(s))

General Options for All Axes in the Layout

The options that are documented in this section can be used with any of the axis types
that are supported within a DATALATTICE or DATAPANEL layout. Subsequent
sections in the chapter document the axis options that are available only for specific axis
types: discrete, linear, log, or time axes. The following table provides a summary of the
general options.

Statement Option Description

ALTDISPLAY Controls which axis features are displayed on

second, fourth, and other even row or column
occurrences of the primary axis.

ALTDISPLAYSECONDARY Controls which features are displayed on

second, fourth, and other even row or column
occurrences of the secondary axis.

DISCRETEOPTS Specifies options for a discrete axis.

DISPLAY Controls which axis features are displayed on

first, third, and other odd row or column
occurrences of the primary axis.

DISPLAYSECONDARY Controls which axis features are displayed on

first, third, and other odd row or column
occurrences the secondary axis.

GRIDATTRS Specifies the attributes of the grid lines.

Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1033

Statement Option Description

GRIDDISPLAY Specifies whether axis grid lines are

displayed.

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the

axis label.

LABELFITPOLICY Specifies a policy for fitting axis labels in the

available space.

LABELPOSITION Specifies the position of the axis label.

LABELSPLITCHAR Specifies one or more characters on which the

axis labels can be split, if needed.

LABELSPLITCHARDROP Specifies whether the split characters should

be included in the axis labels that are
displayed.

LABELSPLITJUSTIFY Specifies the justification of the strings that

are inside the axis label blocks.

LINEAROPTS Specifies options for a standard numeric

interval axis.

LOGOPTS Specifies options for a log axis.

NAME Assigns a name to an axis for reference in

other statements.

OFFSETMAX Reserves an area at the maximum end of the

axis. No tick marks are displayed in the
reserved area.

OFFSETMIN Reserves an area at the minimum end of the

axis. No tick marks are displayed in the
reserved area.

REVERSE Specifies whether the tick values should

appear in the reverse order.

SHORTLABEL Specifies an alternate axis label.

TICKVALUEATTRS Specifies the color and font attributes of the

axis tick value labels.

TICKVALUEHALIGN Specifies the horizontal alignment for all of

the tick values that are displayed on the Y and
Y2 axes.
1034 Chapter 8 • Axis Options in Layouts

Statement Option Description

TICKVALUEVALIGN Specifies the vertical alignment for all of the

tick values that are displayed on the X and X2
axes.

TIMEOPTS Specifies options for a TIME axis.

TYPE Specifies the type of axis to use.

ALTDISPLAY=STANDARD | ALL | NONE | (display-options

controls which axis features are displayed on second, fourth, and other even row or
column occurrences of the primary axis.
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed
NONE
specifies that no axis features are displayed
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:
LABEL
displays the axis label. A common label is displayed at the center of the grid,
and the label applies to all the primary axes in the row or column. This label
is overridden by a label specified on the DISPLAY= option.
LINE
displays the axis line.
TICKS
displays the tick marks.
TICKVALUES
displays the values that are represented by the major tick marks.

Default The settings on the DISPLAY= option.

Tips The default line attributes for the axis line and axis tick marks are defined
in the GraphAxisLine style element.

This option can be used to obtain the alternating axis information as seen in
a ScatterPlotMatrix.

See “Details” on page 1083 for more information about the primary and
secondary axes.

GRIDDISPLAY= and GRIDATTRS= for setting axis grid lines.

ALTDISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which features are displayed on second, fourth, and other even row or
column occurrences of the secondary axis. A secondary axis is not an independent
axis. Rather, it mirrors the primary axis (though it can use different display features).
Thus, for this option to take effect, all plot statements in the LAYOUT PROTOTYPE
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1035

must map data to the same primary axis. For example, a secondary X2 axis can be
displayed on top in the layout, provided all plot statements set XAXIS=X to map
data to the primary X axis (bottom). Similarly, a secondary Y2 axis can be displayed
to the right in the layout, provided all plot statements set YAXIS=Y to map data to
the primary Y axis (left).
NONE
specifies that no axis features are displayed
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed on the
secondary axis
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:
LABEL
displays the axis label. A common label is displayed at the center of the grid,
and the label applies to all the secondary axes in the row or column. This
label is overridden by a label specified on the DISPLAY= option.
LINE
displays the axis line.
TICKS
displays the tick marks.
TICKVALUES
displays the values that are represented by the major tick marks.

Default The settings on the DISPLAYSECONDARY= option.

Restriction If some plot statements set XAXIS=X and others set XAXIS=X2, then
both the X and X2 axis are primary and a secondary X axis cannot be
displayed. In that case, this option is ignored. The same applies for the
Y axes.

Tips The default line attributes for the axis line and axis tick marks are
defined in the GraphAxisLine style element.

This option can be used to obtain the alternating axis information as

seen in a ScatterPlotMatrix.

See “Details” on page 1083 for more information about the primary and
secondary axes.

GRIDDISPLAY= and GRIDATTRS= for setting axis grid lines.

DISCRETEOPTS=(discrete-axis-options)
specifies one or more options for a discrete axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is space
separated.

Interaction This option is ignored if the axis type is not DISCRETE.

1036 Chapter 8 • Axis Options in Layouts

See “Options for Discrete Axes Only” on page 1047 for the options that
you can use for discrete-axis-options.

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on first, third, and other odd row or
column occurrences of the primary axis.
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed
NONE
specifies that no axis features are displayed
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL displays the axis label

LINE displays the axis line
TICKS displays the tick marks
TICKVALUES displays the values that are represented by the major tick
marks

Default STANDARD

Note When LABEL is specified, a common label is displayed at the center of the
grid, and the label applies to all the primary axes in the row or column.
This label overrides a label specified on the ALTDISPLAY= option.

Tips The default line attributes for the axis line and axis tick marks are defined
in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall chart
can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

See “Details” on page 1083 for more information about the primary and
secondary axes..

DISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which axis features are displayed on first, third, and other odd row or
column occurrences of the secondary axis. A secondary axis is not an independent
axis. Rather, it mirrors the primary axis (though it can use different display features).
Thus, for this option to take effect, all plot statements in the LAYOUT PROTOTYPE
must map data to the same primary axis. For example, a secondary X2 axis can be
displayed on top in the layout, provided all plot statements set XAXIS=X to map
data to the primary X axis (bottom). Similarly, a secondary Y2 axis can be displayed
to the right in the layout, provided all plot statements set YAXIS=Y to map data to
the primary Y axis (left).
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1037

NONE
specifies that no axis features are displayed
STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis
ALL
specifies that LABEL, LINE, TICKS, and TICKVALUES are displayed on the
secondary axis
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL displays the axis label

LINE displays the axis line
TICKS displays the tick marks
TICKVALUES displays the values that are represented by the major tick
marks

Default NONE

Restriction If some plot statements set XAXIS=X and others set XAXIS=X2, both
the X and X2 axis are primary and a secondary X axis cannot be
displayed. In that case, this option is ignored. The same applies for the
Y axes.

Note When LABEL is specified, a common label is displayed at the center of

the grid, and the label applies to all the primary axes in the row or
column. This label overrides a label specified on the ALTDISPLAY=
option.

Tip Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis
grid lines.

See for more information about the primary and secondary axes.“Details”
on page 1083 .

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tip On a log axis, this option affects the appearance of the major grid lines
only. It does not affect the appearance of the minor grid lines. To
control the appearance of the minor grid lines on a log axis, use the
MINORGRIDATTRS= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

1038 Chapter 8 • Axis Options in Layouts

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template to
absolutely control the display of grid lines or to allow interaction with the current
style to decide whether grid lines are displayed.
AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element in
the current style contains DisplayOpts="ON."
AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in the
current style contains DisplayOpts="OFF."
ON
specifies that grid lines are always displayed. The current style has no override.
OFF
specifies that grid lines are never displayed. The current style has no override.
The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines style
element. Most supplied templates use the default setting AUTO_OFF to indicate a
preference for not displaying grid lines, but allowing the style to override.

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no

ON any value yes

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style has
no preference about grid lines and the graphics template setting for grid
lines is always used.

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The list
form implies that all included string literals or dynamics will be concatenated.

Default The default label is derived from the primary plot in the layout. For
more information, see “When Plots Share Data and a Common Axis”
on page 880.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1039

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the axis label is too long to fit along the axis, then it is truncated by
default.

Tip Use the SHORTLABEL= option to specify an alternate axis label to be

used whenever truncation would normally occur.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display the axis label.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LABELFITPOLICY=AUTO | SPLIT | SPLITALWAYS

specifies a policy for fitting axis labels in the available space.
AUTO
uses the short label, when specified, instead of the original label. If the short label
does not fit, then it is clipped. When no short label is specified, the original label
is clipped.
SPLIT
splits the axis label at a split character, which is specified by the
LABELSPLITCHAR= option, only when necessary in order to make the label fit
the available space. The short label is not used. A split does not occur at a split
character if a split is not needed at that location. If the label does not contain any
of the specified split characters, then it is not split. A label that cannot be split or
that does not fit the available space even after splitting might overlap the
adjoining space.
SPLITALWAYS
always split the axis label at every occurrence of a split character, which is
specified by the LABELSPLITCHAR= option. If the label cannot be split, then it
is clipped.

Default AUTO

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

When the overlay layout is nested in a lattice layout, SPLIT is ignored

and AUTO is used instead.

Note When LABELPOSITION=CENTER, the available area is the full

axis, including the axis offsets. When
LABELPOSITION=DATACENTER, the available area is the tick
display area, excluding the axis offsets.
1040 Chapter 8 • Axis Options in Layouts

LABELPOSITION=CENTER | DATACENTER | TOP | BOTTOM | LEFT |

RIGHT
specifies the position of the axis label.
CENTER
centers the axis label in the axis area. For the Y and Y2 axes, the label is oriented
vertically and is centered in the axis area (including the offsets). The label is
positioned to the left of the tick values for the Y axis or to the right of the axis
values for the Y2 axis. For the X and X2 axes, the label is centered in the axis
area (including the offsets). It is positioned below the tick values for the X axis or
above the axis values for the X2 axis.
DATACENTER
repeats the axis label for each row or column and centers each label in the axis
tick display area of its row or column. For the Y and Y2 axes, each label is
oriented vertically and is centered in the axis tick display area (excluding the
offsets) of its row. The labels are positioned to the left of the tick values for the Y
axis or to the right of the axis values for the Y2 axis. For the X and X2 axes, each
label is centered in the axis tick display area (excluding the offsets) of its column.
The labels are positioned below the tick values for the X axis or above axis
values for the X2 axis.
TOP | BOTTOM
orients the label horizontally at the top or bottom of the axis area. The label is
right-justified in the axis area for the Y axis and left-justified for the Y2 axis. If
there is not sufficient room in the axis area to display the label, then the label
grows to the right for the Y axis and to the left for the Y2 axis.

Restriction These options are valid for the Y and Y2 axes only.

Note When TOP or BOTTOM is used, the label might collide with other
graphical features.

LEFT | RIGHT
positions the label to the left or right of the axis area. The label is centered
vertically in the axis area.

Restriction These options are valid for the X and X2 axes only.

Note When LEFT or RIGHT is used, the label might collide with other
graphical features.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1041

The following figure shows the CENTER and DATACENTER positions for a blue Y
axis label Qtr and a red X axis label Close.

The next figure shows the TOP and LEFT positions, and the BOTTOM and RIGHT
positions for the same axis labels.

Default CENTER

Restriction This option does not support collision avoidance. In some cases, axis
label collisions can occur in the axis area.

Interaction When LEFT, RIGHT, TOP, or BOTTOM is in effect, the

SHORTLABEL= option is ignored.

See SHORTLABEL= on page 1044 for information about how short labels
are used.

LABELSPLITCHAR="character-list"
specifies one or more characters on which the axis labels can be split, if needed.
When multiple split characters are specified, each character in the list is treated as a
1042 Chapter 8 • Axis Options in Layouts

separate split character unless the specified characters appear consecutively in the
axis label. In that case, all of the specified split characters together are treated as a
single split character.
When LABELFITPOLICY=SPLIT, if the axis label does not fit the available space,
then it is split on a specified split character only if a split is needed at that point to
make the label fit. In this case, a split might not occur on every split character. When
LABELFITPOLICY=SPLITALWAYS, the axis label is split unconditionally on
every occurrence of a split character. If the axis label does not contain any of the
specified split characters, the label is not split.
"character-list"
one or more characters with no space between each character and enclosed in
quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between them.

For example, to specify the split characters a, b, and c, use the
following option:
labelsplitchar="abc"

The LABELSPLIT=TRUE option must also be specified.

Interactions This option has effect only when LABELPOSITION= is CENTER

or DATACENTER.

The LABELSPLITCHARDROP= option specifies whether the split

characters are included in the displayed data label or are dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the LABELSPLITJUSTIFY= option to specify the justification

of the strings in the axis label block.

LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed axis labels.
TRUE
drops the split characters from the axis label display.
FALSE
includes the split characters in the axis label display. When
LABELSPLIT=TRUE and LABELSPLITCHARDROP=FALSE, each split
character remains as the last character in the current line. The characters that
follow the split character, up to and including the next split character, are then
wrapped to the next line.

Default TRUE. The split characters are dropped from the axis label.

Requirement The LABELSPLIT=TRUE option must also be specified.

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1043

The LABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page 1339 for other Boolean values that you can use.

LABELSPLITJUSTIFY=justification
specifies the justification of the strings that are inside the axis label blocks.
justification
CENTER | LEFT | RIGHT
specifies the justification for the X or X2 axis label.
CENTER | TOP | BOTTOM
specifies the justification for the Y or Y2 axis label.

Default CENTER

Requirement LABELSPLIT=TRUE option must also be specified.

Interaction This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

LINEAROPTS=(linear-axis-options)
specifies one or more options for a numeric interval axis. Options must be enclosed
in parentheses. Each option is specified as a name = value pair and each pair is space
separated.

Interaction This option is ignored if the axis type is not LINEAR.

See “Options for Linear Axes Only” on page 1055 for the options that you
can use for linear-axis-options.

LOGOPTS=(log-axis-options)
specifies one or more options for a log axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is space
separated.

Interaction This option is ignored if the axis type is not LOG.

See “Options for Log Axes Only” on page 1065 for the options that you
can use for log-axis-options.

NAME="string"
assigns a name to an axis for reference in other statements. Currently, it is used only
in an AXISLEGEND statement.

Interactions This option is ignored unless the axis is discrete. The axis can be
discrete by default, or explicitly set to discrete with a TYPE=
DISCRETE setting.

For this option to take effect, an axis legend must be enabled. To

enable an axis legend, the DISCRETEOPTS= option must set the
TICKVALUEFITPOLICY to either EXTRACT or
EXTRACTALWAYS. In addition, an AXISLEGEND statement must
be specified to generate the axis legend.

OFFSETMAX=AUTO | AUTOCOMPRESS | number

reserves an area at the maximum end of the axis. No tick marks are displayed in the
reserved area.
1044 Chapter 8 • Axis Options in Layouts

AUTO
reserves just enough area to fully display markers and other graphical features
near the maximum end of an axis.
AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.
number
specifies the offset as a decimal proportion of the full axis length. For a
continuous axis, the offset follows the highest data value or highest tick value,
whichever is greater.

Default AUTOCOMPRESS

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886

OFFSETMIN=AUTO | AUTOCOMPRESS | number

reserves an area at the minimum end of the axis. No tick marks are displayed in the
reserved area.
AUTO
reserves just enough area to fully display markers and other graphical features
near the maximum end of an axis.
AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.
number
specifies the offset as a decimal proportion of the full axis length. For a
continuous axis, the offset precedes the lowest data value or lowest tick value,
whichever is less.

Default AUTOCOMPRESS

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 886.

REVERSE=TRUE | FALSE
specifies whether tick values should appear in the reverse order.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

SHORTLABEL="string"
specifies an alternate axis label to display when the default label or the label
specified by the LABEL= option is too long to fit the available space.
When LABELPOSITION=CENTER (default), the available space for an axis label is
the full axis, including the axis offsets. When LABELPOSITION=DATACENTER,
the available space for an axis label is the axis tick display area, which excludes the
axis offsets. If the label length exceeds the available space, then the label is anchored
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1045

at the left or bottom offset. It extends beyond the opposing offset until it reaches the
end of the axis where it is truncated. An ellipsis designates the truncation.

Interactions This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

This option has effect only when the LABELPOSITION= option is set
to CENTER or DATACENTER.

Note If the specified label is itself too long for the grid length or the grid
width, then it is truncated in the display.

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick values.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TICKVALUEHALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment for all of the tick values that are displayed on the
Y and Y2 axes.

Defaults RIGHT for a Y axis

LEFT for a Y2 axis

Restriction This option is valid for the Y and Y2 axes only.

TICKVALUEVALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment for all of the tick values that are displayed on the X
and X2 axes.
1046 Chapter 8 • Axis Options in Layouts

Defaults TOP for an X axis

BOTTOM for an X2 axis

Restriction This option is valid for the X and X2 axes only.

TIMEOPTS=(time-axis-options)
specifies one or more options for a time axis.

Requirements Columns associated with a time axis must be in SAS time, SAS
date, or SAS datetime units and have an associated SAS time, date,
or datetime format.

Options must be enclosed in parentheses. Each option is specified as

a name = value pair and each pair is space separated.

Interaction This option is ignored if the axis type is not TIME.

See “Options for Time Axes Only” on page 1074 for the options that
you can use for time-axis-options.

TYPE=AUTO | DISCRETE | LINEAR | TIME | LOG

specifies the type of axis to use.
AUTO
requests that the axis type be automatically determined, based on the overlay
contents.
DISCRETE
uses a DISCRETE axis if possible. The data for discrete axes can be character or
numeric. You can add a DISCRETEOPTS= option list to customize this axis
type.
LINEAR
uses a LINEAR axis if possible. You can add a LINEAROPTS= option list to
customize this axis type.
TIME
uses a TIME axis if possible. Data for this axis must be SAS time, SAS date, or
SAS datetime values. You can add a TIMEOPTS= option list to customize this
axis type.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1047

LOG
uses a LOG axis if possible. You can add a LOGOPTS= option list to customize
this axis type.

Interaction If a log axis is requested and the axis data contains 0 or negative
values, the axis reverts to a linear axis. This outcome can occur for
the response axis of a bar chart, line chart, needle plot, or waterfall
chart when a baseline intercept of 0 or less is specified. It can also
occur for the response axis of a waterfall chart when an initial bar
value of 0 or less is specified. To get a log response axis in those
cases, set the baseline intercept or initial bar value to a positive
value.

Default AUTO

Interactions If this option is set to anything other than AUTO, then plots within the
layout are dropped from the display if their data types or data ranges
do not match the axis type requirements. For more information, see
“Plot Axis Types Must Agree on Common Axes” on page 883.

After the axis type is determined (whether you set a specific type or
AUTO is in effect), you can use only options that are supported by
that axis type. For example, if TYPE=TIME, then only the general
OVERLAY axis options and those available on TIMEOPTS= are
supported.

Options for Discrete Axes Only

The options that are documented in this section can be used with the DISCRETEOPTS=
axis option. The following table provides a summary of the options.

Discrete Axis Option Description

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the
TICKVALUELIST= option.

TICKTYPE Specifies the position of the axis tick mark.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the list of tick values that are

displayed on the axis.

TICKVALUEROTATION Specifies how the tick values are rotated on

the X and X2 axes.

TICKVALUESPLITCHAR Specifies a list of characters on which the tick

values can be split, if needed.
1048 Chapter 8 • Axis Options in Layouts

Discrete Axis Option Description

TICKVALUESPLITCHARDROP Specifies whether the split characters are

included in the displayed tick values.

TICKVALUESPLITJUSTIFY Specifies justification of the strings that are

inside the tick value block.

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string values
that are displayed on the axis in place of the values in the TICKVALUELIST=
option. The strings map one-to-one positionally with the values that are listed in the
TICKVALUELIST= option.

Default Determined by the system or by the TICKVALUELIST= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Tip This option should be used with the TICKVALUELIST= option.

The number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

Example The following example specifies the axis tick values 10, 20, 30, and
40, and the tick display values A, B, C, and D:
tickvaluelist=("10" "20" "30" "40");
tickdisplaylist=("A" "B" "C" "D");

TICKTYPE=MIDPOINT | INBETWEEN
specifies the position of the axis tick marks.

MIDPOINT places the tick marks at the midpoint value location.

INBETWEEN places the tick marks half way between adjacent midpoint
locations.

Default MIDPOINT

Restriction This option applies to discrete axes only.

Note Starting with the second maintenance release of SAS 9.4, when
TICKTYPE=INBETWEEN, the outermost tick marks and grid lines at
each end of the axis are not drawn.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness of a
collision-avoidance policy depends on the number of tick values, their length, and
the length of the axis. Which policies are valid depends on the axis on which this
option is used. For the Y and Y2 axes, the following policies are valid:
EXTRACT
displays consecutive integers along the axis instead of the actual tick values in
order to represent those tick values. In most cases, this policy is implemented if
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1049

the system estimates that a collision might occur. If no collision occurs, then the
actual tick values are displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User's Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User's Guide.

NONE
makes no attempt to avoid collisions between tick values. Tick values are display
even when they collide.
SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space. A split does not occur at a split character if a split is
not needed at that location. If the value does not contain any of the specified split
characters, then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR=

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that is
specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR=

SPLITALWAYSTHIN
same as SPLITALWAYS, except that thinning is performed when long words do
not fit the available space.
SPLITTHIN
same as SPLIT, except that thinning is performed when long words do not fit the
available space.
THIN
eliminates alternate tick values.
For the X and X2 axes, the following policies are valid:
EXTRACT
display consecutive integers along the axis instead of the actual tick values to
represent those tick values. In most cases, this policy is implemented if the
system estimates that a collision might occur. If no collision occurs, then the
actual tick values are displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
1050 Chapter 8 • Axis Options in Layouts

Tick Values into a Legend” in SAS Graph Template Language:

User's Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User's Guide.

NONE
does not attempt to fit tick values that collide.
ROTATE
rotates the tick values if a collision occurs. The TICKVALUEROTATION=
option specifies whether the values are rotated to a 45-degree diagonal or a 90-
degree vertical position. By default, the values are rotated to a 45-degree
diagonal position.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs. The
TICKVALUEROTATION= option specifies whether the values are rotated to a
45-degree diagonal or a 90-degree vertical position. By default, the values are
rotated to a 45-degree diagonal position.
ROTATEALWAYSDROP
attempts the ROTATEALWAYS policy, and then drops the tick values if
collisions still occur.
ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space. A split does not occur at a split character if a split is
not needed at that location. If the value does not contain any of the specified split
characters, then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR=

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that is
specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR=

SPLITROTATE
attempts the SPLIT policy, and then the ROTATE policy.
STAGGER
alternates the tick values between two rows.
STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1051

STAGGERTRUNCATE
attempts the STAGGER policy, and then the TRUNCATE policy.
THIN
eliminates alternate tick values.
TRUNCATE
shortens the tick values when they exceed a certain number of characters.
TRUNCATEROTATE
attempts the TRUNCATE policy, and then the ROTATE policy.
TRUNCATESTAGGER
attempts the TRUNCATE policy, and then the STAGGER policy.
TRUNCATETHIN
attempts the TRUNCATE policy, and then the THIN policy.

Defaults ROTATE for the X and X2 axes

THIN for the Y and Y2 axes

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format
specifies how to format the values for major tick marks.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Restrictions This option applies only to discrete axes.

Only character formats are supported.

Interaction This option is ignored when the axis tick values are extracted to an
axis legend. See TICKVALUEFITPOLICY=EXTRACT |
EXTRACTALWAYS on page 1048.

Tip Use this option when you want to duplicate tick values on an axis.

TICKVALUELIST=(string-list)
specifies the list of tick values that are to be displayed on the axis.
string-list
a space-separated list of values, enclosed in parentheses. You must enclose each
value in the list in quotation marks.
Only the tick values that are included in the string list are displayed on the axis. The
values are displayed in the order in which they are listed. The data values that are not
in the list are dropped. The list can be a subset of the data values. It can also contain
values that are not included in the actual data. A tick value that is not included in the
data appears on the axis, but no data is represented at its tick mark.

Requirements The list of values must be enclosed in parentheses.

Each value must be enclosed in quotation marks and separated from

adjacent values by a blank space.

Notes If the string list contains duplicate values, then the first occurrence
of the duplicated value in the list is honored and the remaining
instances are ignored.
1052 Chapter 8 • Axis Options in Layouts

When the values specified in the list are compared with the actual
data values, leading blanks are honored and trailing blanks are
ignored.

Tips You can use this option to subset the axis values or to display the
values in a specific order.

You can use this option to display values on the axis that are not
contained in the data.

Examples The following example specifies the axis tick values Sedan, Sports,
Wagon, and SUV:
tickvaluelist=("Sedan" "Sports" "Wagon" "SUV")

The following example specifies the axis tick values 10, 20, 30, and
40:
tickvaluelist=("10" "20" "30" "40")

TICKVALUEROTATION=DIAGONAL | VERTICAL
specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL rotates the tick values to a 45-degree diagonal position. The X

labels read left to right in a downward direction. The X2 labels
read left to right in an upward direction.
VERTICAL rotates the labels to a 90-degree vertical position. The labels are
always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXISOPTS= and

COLUMN2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESPLITCHAR="character-list"
specifies a list of characters on which the tick values can be split, if needed. When
multiple characters are specified, each character in the list is treated as a separate
split character unless the specified characters appear consecutively in the tick value.
In that case, all of the specified split characters together are treated as a single split
character.
When TICKVALUESPLITPOLICY=SPLIT, if a tick value collision is detected, then
the tick value is split on a split character only if necessary at that point in order to
avoid collision. In that case, a split might not occur on every split character. When
TICKVALUEFITPOLICY=SPLITALWAYS, the tick value is split unconditionally
on every occurrence of a split character. If the tick value does not contain any of the
specified split characters, then it is not split.
"character-list"
one or more characters with no delimiter between each character.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1053

Multiple characters must be specified with no delimiters. For

example, to specify the characters a, b, and c, use the following
option:
tickvaluesplitchar="abc"

Interactions This option is ignored unless option TICKVALUEFITPOLICY= is

set to SPLIT, SPLITALWAYS, SPLITTHIN, or
SPLITALWAYSTHIN.

The TICKVALUEFITPOLICY= option sets the policy that is used

to manage the split behavior of the tick values.

The TICKVALUESPLITCHARDROP= option specifies whether

the split characters are displayed or dropped from the display.

Notes When multiple characters are specified, the order of characters in

the list is not significant.

The split characters are case sensitive.

Tips Use the TICKVALUESPLITJUSTIFY= option to specify the

justification of the strings in the tick value block.

For the X and X2 axis tick values, use the TICKVALUEVALIGN=

option to specify the vertical alignment of the tick values.

For the Y and Y2 axis tick values, use the TICKVALUEHALIGN=

option to specify the horizontal alignment of the tick values.

Example The following example specifies a blank space, a comma, and an

underscore as split characters:
tickvaluesplitchar=" ,_"

TICKVALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters should be included in the displayed tick values.
The split characters are specified by the TICKVALUESPLITCHAR= option.
TRUE
drops the split characters from the tick value display. The following figure shows
an example in which TICKVALUESPLITCHARDROP=TRUE and three-word,
asterisk-delimited tick values are split on the asterisk character by using the
SPLITALWAYS policy.

Notice that the asterisk delimiter is not displayed.

FALSE
includes the split characters in the tick value display. The fit policy determines
how the characters are displayed. If the display policy is SPLIT or SPLITTHIN
and TICKVALUESPLITCHARDROP=FALSE, then each tick value is split at a
split character only where a split is necessary in order to make the value fit the
available space. A split might not occur at every split character in the tick value.
At each split point, the split character remains as the last character in the current
1054 Chapter 8 • Axis Options in Layouts

line. The characters that follow the split character, up to and including the split
character at the next split point, are then wrapped to the following line. This
process repeats until the entire data tick value is displayed. The following figure
shows an example in which TICKVALUESPLITCHARDROP=FALSE and
three-word, asterisk-delimited tick values are split on the asterisk character by
using the SPLIT policy.

Notice that a split occurs on the first asterisk and not at the second. In this case, a
split is not needed at the second asterisk.
If the fit policy is SPLITALWAYS or SPLITALWAYSTHIN, and
TICKVALUESPLITCHARDROP=FALSE, then each tick value is split at every
instance of a split character in the value regardless of whether a split is actually
needed. Each split character remains as the last character in the current line. The
characters that follow each split character, up to and including the next split
character, are then wrapped to the next line. The following figure shows an
example in which TICKVALUESPLITCHARDROP=FALSE and three-word,
asterisk-delimited tick values are split on the asterisk character by using the
SPLITALWAYS policy.

Notice that a split occurs after each asterisk and each asterisk appears at the end
of the line. In this case, three lines are displayed.

Default TRUE

Interactions The TICKVALUESPLITCHAR= option specifies the split character

or characters.

This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUESPLITJUSTIFY=CENTER | LEFT | RIGHT

specifies justification of the strings that are inside the tick value block. The
justification is relative to an individual tick value’s display area and does not affect
the display of tick values that are not split.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1055

Defaults CENTER for an X or X2 axis

RIGHT for a Y axis

LEFT for a Y2 axis

Interaction This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

Options for Linear Axes Only

The options that are documented in this section can be used with the LINEAROPTS=
axis option. The following table provides a summary of the options.

Linear Axis Option Description

INTEGER Specifies that evenly spaced integer values are

used for tick marks.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether the minor tick marks are

displayed on the axis.

THRESHOLDMAX Specifies a bias for including one more tick

mark at the maximum end of the axis.

THRESHOLDMIN Specifies a bias for including one more tick

mark at the minimum end of the axis.

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the
TICKVALUELIST= option.
1056 Chapter 8 • Axis Options in Layouts

Linear Axis Option Description

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision. Only the default policy (THIN) is
available for a Y or Y2 axis.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the order of the tick values for a

linear axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification

can extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on

the X and X2 axes.

TICKVALUESEQUENCE Specifies the tick values for a linear axis by

start, end, and increment.

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the
TICKVALUEFORMAT= option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAMPARM statement.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1057

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
The following figure shows the minor grid lines set to light blue, dotted lines on a linear
axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.
1058 Chapter 8 • Axis Options in Layouts

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Defaults Four ticks with five intervals in the first maintenance release of SAS
9.4 and earlier releases.

One tick with two intervals starting with the second maintenance
release of SAS 9.4.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1059

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 885

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed. If
the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 885

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string values
that are displayed on the axis in place of the values in the TICKVALUELIST=
option. The strings map one-to-one positionally with the values that are listed in the
TICKVALUELIST= option.

Default The display of tick values is controlled by the

TICKVALUEFORMAT= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.
1060 Chapter 8 • Axis Options in Layouts

Interaction When this option is specified, the TICKVALUEFORMAT= option

is ignored.

Tip This option should be used with the TICKVALUELIST= option.

The number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness of a
collision-avoidance policy depends on the number of tick values, their length, and
the length of the axis. Which policies are valid depends on the axis on which this
option is used. For the Y and Y2 axes, the following policies are valid:
NONE
makes no attempt to avoid collisions between tick values. Tick values are display
even when they collide.
THIN
eliminates alternate tick values.
For the X and X2 axes, the following policies are valid:
ROTATE
rotates the tick values if a collision occurs. The TICKVALUEROTATION=
option specifies whether the values are rotated to a 45-degree diagonal or a 90-
degree vertical position. By default, the values are rotated to a 45-degree
diagonal position.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs. The
TICKVALUEROTATION= option specifies whether the values are rotated to a
45-degree diagonal or a 90-degree vertical position. By default, the values are
rotated to a 45-degree diagonal position.
ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
STAGGER
alternates the tick values between two rows.
STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.
THIN
eliminates alternate tick values.

Default THIN

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.
(format-options)
specifies one or more formatting options for major tick values. Together, these
options provide parameters for determining an optimal format (w.d, Ew.,
BESTw.) for displaying major tick values.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1061

MAXWIDTH=integer
specifies the maximum width for displayed tick values. Values might be
rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values. Values
might be rounded or converted to E notation to fit into this width.

Default 6

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values. The actual number might vary based on other constraints.

Default 2

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it to
reduce the tick value width. The scale can be a named scale or a scientific-
notation scale. The EXTRACTSCALETYPE= option specifies the scale type.
The scale that is used is appended to the axis label, as shown in the following
example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the label
is truncated, and the scale is appended to the truncated label. Ellipses indicate
that the label was truncated, as shown in the following example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis. In
that case, the tick value format might differ from the data
format even when a named format is applied to the data values.

See “boolean ” on page 1339 for other Boolean values that you
can use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.
1062 Chapter 8 • Axis Options in Layouts

DEFAULT
extracts a named scale. A named scale can be millions, billions, or
trillions for values of 999 trillion or less, or a multiple of 10 (denoted as
10^n) for values over 999 trillion. For large tick values, the scale factor is
set to ensure that the absolute value of the largest value is greater than 1.
For small fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.
SCIENTIFIC
extracts a scientific-notation scale. A scientific-notation scale is a
multiple of 10 expressed as 10^n for values greater than 1, or a multiple
of 1/10 expressed as 10^–n for values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.

DATA
uses the format that has been assigned to the column that is contributing to the
axis (or BEST6 if no format is assigned) in order to control the formatting of the
major tick values.
format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For details,
see Appendix 4, “SAS Formats Not Supported,” on page 1353.

Note If you specify a format that significantly reduces precision, because

of tick-value rounding, the plot data elements might not align
properly with the axis tick values. In that case, specify a tick-value
format with a higher precision.

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

Interaction This option is ignored when the TICKDISPLAYLIST= option is

specified.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default An internal algorithm determines the tick marks, based on the actual
axis data range or the data range established by the VIEWMIN= and
VIEWMAX= options. By default, when this option is used, the only
tick values that appear are the tick values in numeric-list that fall
within the explicit data range (set by VIEWMIN= and VIEWMAX=)
or the implicit data range (set by the actual data minimum and data
maximum).

Restriction This option applies to linear axes only.

Requirement The tick values must be specified as a space-separated list of numeric

values, enclosed in parentheses.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1063

Interactions This option overrides the INTEGER= option.

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option or the
DISPLAYSECONDARY= option does not display tick values.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then all
ticks in the tick list are displayed. This might result in some data not
being displayed. For example, data might not be displayed when the
VIEWMIN= value is greater than the actual data minimum, or when
the VIEWMAX= value is less than actual data maximum.

If TICKVALUEPRIORITY= TRUE, then the VIEWMIN= and

VIEWMAX= options are ignored if they are fully enclosed by the
numeric-list. The tick numeric-list can extend the implicit data range
of the axis, but cannot reduce it.

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option. If the minimum and maximum of the user-
specified values are within the data range, this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Restriction This option applies to linear axes only.

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | VERTICAL
specifies how the tick values are rotated on the X and X2 axes.
1064 Chapter 8 • Axis Options in Layouts

DIAGONAL rotates the tick values to a 45-degree diagonal position. The X

labels read left to right in a downward direction. The X2 labels
read left to right in an upward direction.
VERTICAL rotates the labels to a 90-degree vertical position. The labels are
always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXISOPTS= and

COLUMN2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESEQUENCE=(sequence-options)
specifies the tick values by start, end, and increment.
(sequence-options)
a space-separated list of the following name-value-pair options that control major
tick values. You must provide all three options.
START=number
specifies the value for the first tick mark.
END=number
specifies the value for the last tick mark.
INCREMENT=number
specifies the increment for intermediate tick marks between the first and last
tick marks. The END value always controls the last tick mark. The interval
between the last tick mark and the previous tick mark might not necessarily
be the INCREMENT value.

Default An internal algorithm determines the tick marks, based on the actual
axis data range or the data range established by the VIEWMIN= and
VIEWMAX= options. By default, when this option is used, the only
tick values that appear are those that fall within the explicit data range
(set by VIEWMIN= and VIEWMAX=) or the implicit data range (set
by the actual data minimum and data maximum).

Interactions This option overrides the INTEGER= option.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the START= option value and the
VIEWMAX= option is set to the END= option value, then all ticks in
the tick sequence are displayed.

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display tick marks.

Tip The values in the sequence are formatted according to the setting for
the TICKVALUEFORMAT= option.

See TICKVALUELIST= option as an alternative for customizing tick

marks.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1065

VIEWMAX=number
specifies the maximum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The maximum value in the data for the specified axis.

Interactions This option does not determine the maximum axis tick value that is
displayed. The THRESHOLDMAX= value is used to determine the
maximum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

VIEWMIN=number
specifies the minimum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The minimum value in the data for the specified axis.

Interactions This option does not determine the minimum axis tick value that is
displayed. The THRESHOLDMIN= value is used to determine the
minimum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Options for Log Axes Only

The options that are documented in this section can be used with the LOGOPTS= axis
option. The following table provides a summary of the options.
1066 Chapter 8 • Axis Options in Layouts

Log Axis Option Description

BASE Specifies the base of the logarithmic scale for

the axis values.

MINORGRID Specifies whether grid lines are displayed at

the minor tick marks.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether minor ticks are displayed.

TICKINTERVALSTYLE Specifies how to scale and format the values

for major tick marks.

TICKVALUEFORMAT= Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the tick values for a log axis as a

space-separated list.

TICKVALUEPRIORITY Specifies whether the TICKVALUELIST

specification can extend the axis data range.

VALUESTYPE Specifies the scale that the system uses when

interpreting the TICKVALUELIST=,
VIEWMAX=, and VIEWMIN= option values.

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

BASE=10 | 2 | E
specifies the base of the logarithmic scale for the axis values.

Default 10

Restriction This option applies to log axes only.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1067

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
The following figure shows the minor grid lines set to light blue, dotted lines on a
base-10 log axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.
1068 Chapter 8 • Axis Options in Layouts

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Default Eight ticks with nine intervals (BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT. ).

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKINTERVALSTYLE=AUTO | LOGEXPAND | LOGEXPONENT | LINEAR

specifies how to scale and format the values for major tick marks.
AUTO
selects a LOGEXPAND, LOGEXPONENT, or LINEAR representation
automatically based on the range of the data. When the data range is small
(within an order of magnitude), a LINEAR representation is typically used. Data
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1069

ranges that encompass several orders of magnitude typically use the

LOGEXPAND or LOGEXPONENT representation.
LOGEXPAND
places the major tick marks at uniform intervals at integer powers of the base.
The tick values are expanded as follows:

Base=10

Base=2

Base=E

LOGEXPONENT
places the major tick marks at uniform intervals at integer powers of the base.
The tick values are only the integer exponents for all bases.

LINEAR
places the major tick marks at non-uniform intervals that cover the range of the
data.

Default AUTO

Restrictions This option applies to log axes only.

For LOGEXPONENT, formats on data columns contributing to the

axis are ignored. For LOGEXPAND, formats on data columns
contributing to the axis are ignored, although any "named format" on
the column is retained. For LINEAR, ticks values are automatically
formatted when the column format is not assigned or one of w.d, Ew.,
or BESTw. Other formats (SAS defined or user-defined) are used if
specified.

GTL currently honors most but not every SAS format. For details, see
Appendix 4, “SAS Formats Not Supported,” on page 1353.

Note When BASE=10 and LOGEXPAND or LOGEXPONENT is used, an

intermediate tick is added whenever the axis data range is less than or
equal to 1.5 powers of 10.

Tip If you use TICKINTERVALSTYLE=LOGEXPONENT, then you

might want to include information in the axis label about which base
is used.

TICKVALUEFORMAT=DATA | format
specifies how to format the values for major tick marks.
1070 Chapter 8 • Axis Options in Layouts

Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
DATA
uses the format that has been assigned to the column that is contributing to the
axis (or BEST6 if no format is assigned) in order to control the formatting of the
major tick values.
format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For details,
see Appendix 4, “SAS Formats Not Supported,” on page 1353.

Restriction This option applies to log axes only.

Interactions This option is ignored when

TICKINTERVALSTYLE=LOGEXPONENT.

When TICKINTERVALSTYLE=LOGEXPAND, this option is

honored for the base 10 and base 2 logarithmic scales, and is ignored
for the base E scale.

When TICKINTERVALSTYLE=LINEAR, this option is honored for

the base 10, base 2, and base E logarithmic scales.

See BASE=

TICKINTERVALSTYLE=

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default Only the tick values specified in the list that fall within the explicit
data range set by the VIEWMIN= and VIEWMAX= options or by
the implicit data range set by the actual data minimum and data
maximum are displayed. An internal algorithm determines the tick
marks.

Requirements The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

The values that you specify must be appropriate for the

VALUESTYPE= specification. Otherwise, unexpected results might
occur. If VALUESTYPE=EXPANDED is in effect (default), specify
increments of the log base power such as 0.1, 1, 10, 100, and so on,
on a base 10 log axis, for example. If VALUESTYPE=EXPONENT
is in effect, specify integer increments of the log base power
exponent such as 1, 2, 3, and so on.

Interactions The VALUESTYPE= option determines how the values in the list
are interpreted.

The VIEWMIN= and VIEWMAX= options alter the axis data

range. If the VIEWMIN= option is set to the minimum tick list
value and the VIEWMAX= option is set to the maximum tick list
value, then all ticks in the tick list are displayed. This might result in
some data not being displayed. For example, data might not be
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1071

displayed when the VIEWMIN= value is greater than the actual data
minimum, or when the VIEWMAX= value is less than actual data
maximum.

If the VIEWMIN= value is greater than the actual data minimum or

the VIEWMAX= value is less than actual data maximum, some data
might not be displayed.

This option is ignored if the DISPLAY= or the

DISPLAYSECONDARY= option does not display the tick values.

See VIEWMIN= and VIEWMAX= options for controlling the data

range

TICKINTERVALSTYLE= for specifying the scale and format of the

major tick values

TICKVALUEPRIORITY= for controlling the behavior of the

TICKVALUELIST= option

BASE= for specifying the log base

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by the TICKVALUELIST= option. If the
minimum and maximum of the user-specified values are within the data range,
this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

VALUESTYPE=EXPANDED | EXPONENT
specifies the scale that the system uses when interpreting the TICKVALUELIST=,
VIEWMAX=, and VIEWMIN= option values. Use this option to choose your
preferred way of specifying log-axis values.
1072 Chapter 8 • Axis Options in Layouts

EXPANDED values are interpreted as powers of the base such as 0.1, 1, 10,
100, and so on, for base 10, for example.
EXPONENT values are interpreted as integer exponents of the base such as 1,
2, 3, and so on, for base 10, base 2, and base E.

Default EXPANDED

Note This option does not change the style of the tick values that are
displayed on the axis. It changes only how the VIEWMIN=,
VIEWMAX=, and TICKVALUELIST= option values are interpreted by
the system.

Tip This option is particularly useful when BASE=E.

Examples The following example specifies VIEWMIN= and VIEWMAX= as

exponent values instead of as expanded values on an expanded Base 10
log axis. This results in X-axis tick values of 10, 100, 1000, 10000, and
100000.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexpand
valuestype=exponent
viewmin=1 viewmax=5));

The following example specifies TICKVALUELIST= as a list of

expanded values instead of exponent values on an exponent Base 10 log
axis. This results in X-axis tick values of 1, 2, 3, 4, and 5.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexponent
tickvaluepriority=true
valuestype=expanded
tickvaluelist=(10 100 1000 10000 100000)));

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Requirement The value that you specify must be appropriate for the
VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is in
effect (default), specify an increment of the log base power such as
0.1, 1, 10, 100, and so on, on a base 10 log axis, for example. If
VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

If an invalid value is specified for the VIEWMAX= option, the

default value for VIEWMAX= is used instead. In that case, if the
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1073

default value for VIEWMAX= is less than the value specified by the
VIEWMIN= option, then the VIEWMIN= and VIEWMAX= values
are swapped.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to 1.5
powers of 10.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Examples The following example specifies a value of 100,000 as an expanded

value on a base 10 log axis:
VIEWMAX=100000

The following example specifies a value of 100,000 as an exponent

value on a base 10 log axis:
VIEWMAX=5

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Requirement The value that you specify must be appropriate for the
VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is in
effect (default), specify an increment of the log base power such as
0.1, 1, 10, 100, and so on, on a base 10 log axis, for example. If
VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the plot
statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to 1.5
powers of 10.
1074 Chapter 8 • Axis Options in Layouts

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Examples The following example specifies a value of 0.1 as an expanded value

on a base 10 log axis:
VIEWMIN=0.1

The following example specifies a value of 0.1 as an exponent value

on a base 10 log axis:
VIEWMIN=-1

Options for Time Axes Only

The options that are documented in this section can be used with the TIMEOPTS= axis
option. Some of the options are axis-specific. The following tables provide a summary of
the time-axes options based on the layout axis option in which it is used.

Table 8.4 Time-Axis Options for the COLUMNAXIS= or COLUMN2AXIS=Layout Option

Time Axis Option Description

INTERVAL Specifies that evenly spaced integer values are

used for tick marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time

interval that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKINTERVAL Specifies the time interval between minor

ticks.

MINORTICKS Specifies whether minor ticks are displayed.

SPLITTICKVALUE Specifies whether to split the tick values on an

X or X2 axis, if possible.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value

collision on an X or X2 axis.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the order of the tick values for a

time axis as list.

TICKVALUEROTATION Specifies how the tick values are rotated on

the X and X2 axes.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1075

Time Axis Option Description

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

Table 8.5 Time-Axis Options for the ROWAXIS= or ROW2AXIS=Layout Option

Time Axis Option Description

INTERVAL Specifies that evenly spaced integer values are

used for tick marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time

interval that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at

the minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKINTERVAL Specifies the time interval between minor

ticks.

MINORTICKS Specifies whether minor ticks are displayed.

TICKVALUEFORMAT Specifies how to format the values for major

tick marks.

TICKVALUELIST Specifies the order of the tick values for a

time axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification

that is specified with the TICKVALUELIST=
option can extend the axis data range.

VIEWMAX Specifies the maximum data value to include

in the display.

VIEWMIN Specifies the minimum data value to include

in the display.

INTERVAL=interval
specifies the time interval between major ticks. Valid interval keywords are AUTO,
SECOND, MINUTE, HOUR, DAY, TENDAY, WEEK, SEMIMONTH, MONTH,
QUARTER, SEMIYEAR, YEAR.
1076 Chapter 8 • Axis Options in Layouts

Table 8.6 Time Intervals

Default tick value

INTERVAL Unit Tick interval format

AUTO DATE, TIME, or automatically automatically

DATETIME chosen chosen

SECOND TIME or second TIME8.

DATETIME

MINUTE TIME or minute TIME8.

DATETIME

HOUR TIME or hour TIME8.

DATETIME

DAY DATE or day DATE9.

DATETIME

TENDAY DATE or 10 days DATE9.

DATETIME

WEEK DATE or 7 days DATE9.

DATETIME

SEMIMONTH DATE or 1st and 16th of each DATE9.

DATETIME month

MONTH DATE or month MONYY7.

DATETIME

QUARTER DATE or 3 months YYQC6.

DATETIME

SEMIYEAR DATE or 6 months MONYY7.

DATETIME

YEAR DATE or year YEAR4.

DATETIME

Default AUTO. An appropriate interval is chosen based on the data and the
column date, date-time, or time format.

Restriction This option applies to time axes only.

Requirement The data column(s) mapped to a time axis must be in the same
duration units: TIME, DATE, or DATETIME. The selection of an
interval must be consistent with the duration unit. For example, if the
data are in time units, you can specify only AUTO, SECOND,
MINUTE, HOUR.

Interaction This option is ignored if the TICKVALUELIST= option is used.

Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1077

INTERVALMULTIPLIER=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the axis.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default 1

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= option is used.

Tip Use the INTERVAL= option to specify a different time interval.

Examples To specify 3-month intervals:

INTERVAL=MONTH INTERVALMULTIPLIER=3

To specify 10-year intervals:

INTERVAL=YEAR INTERVALMULTIPLIER=10

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

the second maintenance release of SAS 9.4. If attribute DisplayOpts is
not defined in the active style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the minor
grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page 1339 for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the major
grid lines.
1078 Chapter 8 • Axis Options in Layouts

The following figure shows the minor grid lines set to light blue, dotted lines on a time
axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with the

second maintenance release of SAS 9.4.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR, LINESTYLE, and LINETHICKNESS attributes
are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Line Options” on page 1349 for available line options.

Example Here is an example that specifies light blue, dotted lines for the minor
grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKINTERVAL=interval
specifies the time interval between minor ticks. See Table 8.6 on page 1076 for
information about the intervals that you can select. The interval that you select must
be consistent with the axis data duration units such as TIME, DATE, or DATETIME.
For example, if the axis data is in TIME units, then you must specify AUTO,
SECOND, MINUTE, or HOUR.

Default AUTO

Interactions This option is ignored if the TICKVALUELIST= option is used.

This option is ignored if the MINORTICKINTERVAL= setting is

greater than the INTERVAL= setting.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the minor
tick marks are displayed on the axis as shown in the following figure.
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1079

Default FALSE

Interactions The number of minor ticks is dependent on the value of the

MINORTICKINTERVAL= option, if specified. If
MINORTICKINTERVAL= is not specified, then it is dependent on the
value of the INTERVAL= option.

This option is ignored if the TICKVALUELIST= option is used, or if

the DISPLAY= or DISPLAYSECONDARY= option does not display
the tick marks.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page 1339 for other Boolean values that you can use.

SPLITTICKVALUE=TRUE | FALSE
specifies whether to split the tick values on an X or X2 axis, if possible. This option
is not available for a Y or Y2 axis.
TRUE
splits the axis tick values into two lines allowing more tick values to appear. For
example, with INTERVAL= MONTH, this is how tick values are split:

FALSE
does not split the axis tick values. For example, when this option specifies
FALSE, this is how the tick values in the previous example appear:

Typically, fewer tick values fit, causing thinning, rotation, or staggering of the
values. See the TICKVALUEFITPOLICY= option.

Default TRUE

Restrictions This option applies to time axes only.

This option is valid only in the COLUMNAXIS= and

COLUMN2AXIS= layout axis options.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUEFORMAT= option is used.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an X or X2 axis. This option is
not available for the Y and Y2 axes. The effectiveness of a collision-avoidance
policy depends on the number of tick values, their length, and the length of the axis.
The following policies are valid:
1080 Chapter 8 • Axis Options in Layouts

NONE
makes no attempt to avoid collisions between tick values. Tick values are display
even when they collide.
ROTATE
rotates the tick values if a collision occurs. The TICKVALUEROTATION=
option specifies whether the values are rotated to a 45-degree diagonal or a 90-
degree vertical position. By default, the values are rotated to a 45-degree
diagonal position.
ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs. The
TICKVALUEROTATION= option specifies whether the values are rotated to a
45-degree diagonal or a 90-degree vertical position. By default, the values are
rotated to a 45-degree diagonal position.
ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.
STAGGER
alternates the tick values between two rows.
STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.
STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.
THIN
eliminates alternate tick values.

Default THIN

Restriction This option is valid only for the X and X2 axes.

Interaction When SPLITTICKVALUE= TRUE, this option is ignored and only the
THIN policy is used.

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format | DATA
specifies how to format the values for major tick marks.
format
specifies a SAS date, time, or datetime format to control how the major tick
values are displayed. This format must be in the same duration units as the data
column(s) mapped to a time axis: TIME, DATE, or DATETIME and should be
appropriate for the value of the INTERVAL= option. For example, if
INTERVAL=MONTH and there are two years of data displayed on the axis, then
choosing TICKVALUEFORMAT=YEAR. would result in several ticks having
the same year value.
DATA
specifies that the SAS date, time, or datetime format associated with the data
column assigned to the axis be used to control how the major tick values are
displayed.

Default The default format used by the INTERVAL= option. The default does
not apply if TICKVALUELIST= is specified.

Restrictions This option applies to time axes only.

Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1081

GTL currently honors most but not every SAS format. For details, see
Appendix 4, “SAS Formats Not Supported,” on page 1353.

TICKVALUELIST=(time-constant-list | date-constant-list | datetime-constant-list |

numeric-list)
specifies the tick values for a time axis as list.

Default An internal algorithm determines the tick values.

Restrictions This option applies to time axes only.

If TICKVALUEPRIORITY= is set to FALSE, then this option does

not extend the data range of the axis. If the values fall within the
default data range or that specified by the VIEWMIN= or
VIEWMAX= options, then they are used.

Requirement The tick values must be specified as a space-separated list of values

enclosed in parentheses. The items in the list must be in the same
duration units as the data mapped to the axis: TIME, DATE, or
DATETIME. The values can be expressed as SAS TIME, DATE, or
DATETIME constants (for example, "13:23"T, "11MAY06"D, or
"11MAY06:13:23"DT) or their numeric equivalents.

Interactions The values in the list are formatted according to the format specified
on the TICKVALUEFORMAT= option. If TICKVALUEFORMAT=
is not used, then the values are formatted according to the column
format (the default TICKVALUEFORMAT value is not applied to
these values).

If this option is specified, the SPLITTICKVALUE= and

INTERVAL= options are ignored.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.
TRUE
extends the axis data range (but does not reduce it) to include the minimum and
maximum values that are specified by the TICKVALUELIST= option. If the
minimum and maximum of the user-specified values are within the data range,
this option has no effect.
FALSE
displays only the tick values that are specified by the TICKVALUELIST= option
that fall within the explicit data range set by the VIEWMIN= and VIEWMAX=
options or by the implicit data range set by the actual data minimum and data
maximum.

Default FALSE

Restriction This option is valid only in the ROWAXIS= and ROW2AXIS= layout
axis options.

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.
1082 Chapter 8 • Axis Options in Layouts

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page 1339 for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | VERTICAL
specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL rotates the tick values to a 45-degree diagonal position. The X

labels read left to right in a downward direction. The X2 labels
read left to right in an upward direction.
VERTICAL rotates the labels to a 90-degree vertical position. The labels are
always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXISOPTS= and

COLUMN2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the original
data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX= value.
The VIEWMIN= and VIEWMAX= values, and additional factors such as
thresholds and the tick values computed by the plot statement, are used to
determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the original
data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN= value. The
VIEWMIN= and VIEWMAX= values, and additional factors such as
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1083

thresholds and the tick values computed by the plot statement, are used to
determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use the
TICKVALUELIST= option.

See “Adjusting the Axis View” on page 884

Details
The LAYOUT DATALATTICE and LAYOUT DATAPANEL statements each create a
grid of graphs based on the values of one or more classification variables. In the grid, the
axes are always external and displayed on the “primary” axes by default. The axes that
are considered primary depend on the settings for the XAXIS= and YAXIS= options in
plot statements that are specified within the LAYOUT PROTOTYPE. For managing the
primary axes, both the LAYOUT DATALATTICE and LAYOUT DATAPANEL
statements have COLUMNAXISOPTS=, COLUMN2AXISOPTS=, ROWAXISOPTS=,
and ROW2AXISOPTS= options that manage the axis features separately for columns
and rows. The settings that are available can manage odd and even columns and rows
separately, enabling you to simplify the axis display within the grid.
The following table shows which axis is primary for the XAXIS= and YAXIS= settings,
and which axis option to use to manage that primary axis.

Option Setting Primary Axis Axis Option to Use

XAXIS=X X (bottom) COLUMNAXISOPTS=

XAXIS=X2 X2 (top) COLUMN2AXISOPTS=

YAXIS=Y Y (left) ROWAXISOPTS=

YAXIS=Y2 Y2 (right) ROW2AXISOPTS=

The settings that are available for the axis options can manage odd and even columns
and rows separately, enabling you to simplify the axis display within the grid.
• To manage the first, third, and odd occurrences of a primary axis, use the DISPLAY=
option.
• To manage the second, fourth, and even occurrences of a primary axis, use the
ALTDISPLAY= option.
You can also display “secondary” axes in the grid. A secondary axis is not an
independent axis. Rather, it mirrors the primary axis, but it is displayed on the opposite
side and can have different display options. For example, when the X axis (bottom) is
primary, you can mirror that axis with a secondary X axis at the top of the grid.
Similarly, when the Y2 axis (right) is primary, you can mirror that axis with a secondary
Y2 axis on the left of the grid. A secondary axis makes it easier to interpolate values in
the cells that are farthest away from the primary axis.
Secondary axes can be displayed in the graph, provided all plot statements in the
LAYOUT PROTOTYPE map data to the same primary axis. For example, a secondary
X axis can be displayed at the top of the layout, provided all plot statements set
XAXIS=X to map data to the primary X axis (bottom). Similarly, a secondary Y2 axis
can be displayed to the left in the layout, provided all plot statements set YAXIS=Y2 to
map data to the primary Y2 axis (right). If all plot statements in the LAYOUT
1084 Chapter 8 • Axis Options in Layouts

PROTOTYPE do not map data to the same primary axis, then the secondary axes are not
displayed.
To display secondary axes in the grid, use the DISPLAYSECONDARY= and
ALTDISPLAYSECONDARY= options. As with the options for the primary axes, the
DISPLAYSECONDARY= option manages the first, third, and odd occurrences of a
secondary axis. The ALTDISPLAYSECONDARY= option manages the second, fourth,
and even occurrences of a secondary axis.
In the default cases for the plots within the layout, the axis type is always DISCRETE,
LINEAR, or TIME. The TYPE= option enables you to specify an axis type that
overrides the default. For example, when appropriate for the data, you can request a
LOG axis. When you override the default axis type, you must be sure to specify the
correct axis type for the plot(s) that you are defining.
Each axis type has features specific to that type, and the following axis options enable
you to specify features for the different types: DISCRETEOPTS= , LINEAROPTS= ,
LOGOPTS= , and TIMEOPTS= . One or more of these options can be specified for an
axis, but the specified settings are applied only to the axis type that supports them.
Note: Certain plot types or layouts sometimes impose restrictions on what type of axis
can be assigned. See the plot or layout documentation for default axis types and any
restrictions that might apply.

Example: Axis Options for LAYOUT

DATALATTICE and LAYOUT DATAPANEL

The following graph was generated by the “Example Program” on page 1085:
Example: Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1085

Example Program
This example shows how axis attributes can be managed separately for even and odd
columns and rows in the layout grid. In this case, the ROWAXISOPTS= option is used
to stagger the Y-axes attributes. On the primary (left) Y axis, DISPLAY= displays
TICKS and TICKVALUES on the first and third rows while ALTDISPLAY= displays
just TICKS on the second row. On the secondary (right) Y axis,
DISPLAYSECONDARY= displays just TICKS on the first and third rows, while
ALTDISPLAYSECONDARY= displays TICKS and TICKVALUES on the second row.
This alternating pattern could also have been set for the column axes. The pattern is
independent of the number of rows and columns.
proc template;
define statgraph layoutdatalattice;
begingraph;
entrytitle "Annual Furniture Sales Comparisons";
layout datalattice rowvar=country columnvar=year /
rowdatarange=union
headerlabellocation=inside
headerlabeldisplay=value
headerbackgroundcolor=GraphAltBlock:color
rowaxisopts=(griddisplay=on
display=(tickvalues)
altdisplay=(ticks)
displaysecondary=(ticks)
altdisplaysecondary=(ticks tickvalues)
linearopts=(tickvalueformat=dollar12.))
columnaxisopts=(display=(tickvalues)
timeopts=(tickvalueformat=monname3.));
layout prototype / cycleattrs=true;
seriesplot x=month y=TotalActual / name="Actual";
seriesplot x=month y=TotalPredict / name="Predict";
endlayout;
sidebar / align=bottom;
discretelegend "Actual" "Predict" / border=false;
endsidebar;
endlayout;
endgraph;
end;
run;

proc summary data=sashelp.prdsal2 nway;

class country year month;
var actual predict;
output out=prdsal2 sum=TotalActual TotalPredict;
run;

proc sgrender data=prdsal2 template=layoutdatalattice;

run;
1086 Chapter 8 • Axis Options in Layouts
1087

Part 6

Legend Statements

Chapter 9
Legend Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1089
1088
1089

Chapter 9
Legend Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1089
AXISLEGEND Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1089
CONTINUOUSLEGEND Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1098
DISCRETELEGEND and MERGEDLEGEND Statements . . . . . . . . . . . . . . . . 1109
LEGENDITEM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1126
LEGENDTEXTITEMS Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1131

Dictionary

AXISLEGEND Statement
Generates consecutive integers for display as axis-tick values in the graph, and creates a legend that
correlates those integers with the actual tick values that they represent.
Requirements: The axis must be named with the axis NAME= option.
The axis must be a discrete axis (TYPE=DISCRETE).
The axis must use the TICKVALUEFITPOLICY=EXTRACT or
TICKVALUEFITPOLICY=EXTRACTALWAYS suboption of the DISCRETEOPTS=
axis option.
Interaction: A legend might be dropped if the total legend area in the graph exceeds the
percentage that is set by the MAXLEGENDAREA= option in an ODS GRAPHICS
statement that is in effect for the output destination. A legend might also be dropped
if DISPLAYCLIPPED= FALSE and the full legend cannot be displayed.

Syntax
AXISLEGEND "axis-name" </option(s)>;

Summary of Optional Arguments

Appearance options
ACROSS=positive-integer
specifies the number of legend entries that are placed horizontally before the
next row begins.
BACKGROUNDCOLOR=style-reference | color
1090 Chapter 9 • Legend Statements

specifies the color of the legend background.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the legend.
DISPLAYCLIPPED=TRUE | FALSE
specifies whether the legend is displayed when any portion of the legend
cannot be rendered entirely.
DOWN=positive-integer
specifies the number of legend entries that are placed vertically before the
next column begins.
OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or transparent
(FALSE).
ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether legend entries are organized into rows or into columns.
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space that is added outside the legend border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend
perimeter.
VALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend values.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the legend is automatically aligned within its parent layout
when nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies the horizontal alignment of the legend within its parent layout when
nested within an overlay-type or region layout.
LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when the
legend is specified within an overlay-type or region layout.
VALIGN=CENTER | TOP | BOTTOM | number
specifies the vertical alignment of the legend within its parent layout when
nested within an overlay-type or region layout.

Text options
TITLE="string"
specifies the title of the legend.
TITLEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend title.

Required Argument
"axis-name"
references an axis’s name. The name controls which axis determines the legend
entries.

Requirement The axis name must be enclosed in quotation marks.

AXISLEGEND Statement 1091

Optional Arguments
ACROSS=positive-integer
specifies the number of legend entries that are placed horizontally before the next
row begins.

Default The entries are placed to best fit the available area. This “best fit”
approach works only when the legend is nested in the template’s
outermost layout.

Interaction This option is ignored if ORDER= COLUMNMAJOR

AUTOALIGN=NONE | AUTO | (location-list)

specifies whether the legend is automatically aligned within its parent layout when
nested within an overlay-type layout.
NONE
does not automatically align the legend within its parent layout. The legend’s
position is set by the HALIGN= and VALIGN= options.
AUTO
within the parent layout, attempts to center the legend in the area that is farthest
from any surrounding data point markers.
(location-list)
within the parent layout, restricts the legend’s possible locations to those
locations in the specified location-list, and use the location-list position that least
collides with the parent layout’s other graphics features. The location-list is
space-separated and can contain any of these locations: TOPLEFT, TOP,
TOPRIGHT, LEFT, CENTER, RIGHT, BOTTOMLEFT, BOTTOM, and
BOTTOMRIGHT.

Default NONE

Restriction AUTO is available only when the parent layout contains a scatter plot.
Otherwise, it is ignored.

Interactions This option has no effect unless LOCATION= INSIDE.

When LOCATION=INSIDE and AUTOALIGN= is not NONE, the

HALIGN= and VALIGN= options are ignored.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type layout.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphLegendBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect for the color to be seen. By default,
OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.
1092 Chapter 9 • Legend Statements

Defaults TRUE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphLegendBackground:FrameBorder style reference starting with

the second maintenance release of SAS 9.4. If attribute FrameBorder is
not defined in the active style, then TRUE is the default value.

Tip The BORDERATTRS= option controls the appearance of the legend

border.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the legend.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element

“Line Options” on page 1349 for available line-options.

DISPLAYCLIPPED=TRUE | FALSE
specifies whether the legend is displayed when any portion of the legend cannot be
rendered entirely. Based on the legend contents and placement, or when the graph
size is reduced, parts of the legend (title, legend symbol, or legend value) might be
clipped (truncated). When clipping occurs and this option is set to FALSE, the entire
legend is removed from the graph. The space for that legend is then reclaimed by the
remainder of the graph. When this option is set to TRUE, the legend always appears,
even if some parts of it have been clipped.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

DOWN=positive-integer
specifies the number of legend entries that are placed vertically before the next
column begins.

Default The entries are placed to best fit the available area. This “best fit”
approach works only when the legend is nested in the template’s
outermost layout.

Restriction This option is ignored if ORDER= ROWMAJOR

HALIGN=CENTER | LEFT | RIGHT | number

specifies the horizontal alignment of the legend within its parent layout when nested
within an overlay-type or region layout.
number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s width, where 0 is all
the way to the left and 1 is all the way to the right.
AXISLEGEND Statement 1093

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER

Restriction This option is available only when this statement is nested within an
overlay-type or region layout.

Interactions If LOCATION= OUTSIDE, then the HALIGN= and VALIGN=

options cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the

AUTOALIGN= option is not NONE, and the parent layout is an
overlay-type layout.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type or region layout.

LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when the legend
is specified within an overlay-type or region layout.

Default OUTSIDE

Restriction This option has effect only when the legend statement appears within
an overlay-type or region layout and at least one stand-alone plot
statement is referenced by the parent layout.

Interactions The actual position is determined by the settings for the LOCATION=,
AUTOALIGN= , HALIGN= , and VALIGN= options.

If this option is set to OUTSIDE, then the HALIGN= and VALIGN=

options must specify a keyword (LEFT, RIGHT, or CENTER). The
number setting for the alignment is invalid when the legend is
positioned outside of the plot area.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type or region layout..

OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether legend entries are organized into rows or into columns.

Default ROWMAJOR
1094 Chapter 9 • Legend Statements

Interaction If ORDER=ROWMAJOR, then use the ACROSS= option to limit the

number of entries in a row. If ORDER=COLUMNMAJOR, then use
the DOWN= option to limit the number of entries in a column.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space that is added outside the legend border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the legend border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend perimeter.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the legend perimeter.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
AXISLEGEND Statement 1095

BOTTOM=dimension specifies the amount of extra space added to the

bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default Padding for all sides is 0.

Note The default units for dimension are pixels.

See “dimension” on page 1340

TITLE="string"
specifies the title of the legend.

Default No title

Requirement string must be enclosed in quotation marks.

Interaction When the title is placed to the left of the legend, if

TITLEBORDER=TRUE is in effect, no separator is displayed
between the title and the legend. If TITLEBORDER=FALSE is in
effect in that case, a separator is displayed.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend title.

Default The GraphLabelText style element.

Interactions For this option to have any effect, the TITLE= option must also be
specified.

If one or more text options are specified and they do not include all the
font properties (color, family, size, weight, style), then non-specified
properties will be derived from the GraphLabelText style element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

VALIGN=CENTER | TOP | BOTTOM | number

specifies the vertical alignment of the legend within its parent layout when nested
within an overlay-type or region layout.
number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s height, where 0 is on
the bottom and 1 is on the top.
1096 Chapter 9 • Legend Statements

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. Zero corresponds to

the containing layout's bottom edge plus half the legend height.
Similarly, one corresponds to the containing layout's top edge minus
half the legend height.

Default CENTER

Restriction This option is available only when this statement is nested within an
overlay-type or region layout. It is ignored if the parent layout is not
an overlay-type or region layout.

Interactions If LOCATION= OUTSIDE, then the VALIGN= and HALIGN=

options cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the

AUTOALIGN= option is not NONE, and the parent layout is an
overlay-type layout.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type or region layout.

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend values.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all the
font properties (color, family, size, weight, style), then non-specified
properties will be derived from the GraphLabelText style element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

Details
An AXISLEGEND statement is used to consolidate the axis display so that a large
number of tick values can be clearly displayed without collision. In an axis legend, the
legend symbol is a positive integer that is generated to represent a specific axis tick
position, and the legend value displays the axis tick value. In the “Example:
AXISLEGEND Statement” on page 1097 , the first legend symbol is the integer 1 and
the first legend value is Neon SE 4dr.
To implement an axis legend, all of the following requirements must be met:
Example: AXISLEGEND Statement 1097

• The axis must be a discrete axis. The axis can be discrete by default, or explicitly set
to discrete by setting TYPE=DISCRETE among the parent layout’s axis-option
settings.
• The parent layout’s axis options must specify NAME= to assign a name to the axis.
• The parent layout’s axis options must use the DISCRETEOPTS= option to specify
either TICKVALUEFITPOLICY=EXTRACT or
TICKVALUEFITPOLICY=EXTRACTALWAYS. EXTRACT implements an axis
legend if the system estimates that a collision might occur. EXTRACTALWAYS
implements an axis legend regardless of whether a collision occurs.
• The AXISLEGEND statement must reference the axis name that was specified in the
parent layout’s NAME= axis option.
When all of these requirements are satisfied, the tick values of the named discrete axis
can be replaced with consecutive positive integers. The axis legend displays those
integer values and the corresponding tick values that the integers represent.
Within an overlay-type layout, when an axis legend is placed inside the plot area with
LOCATION= INSIDE,
• The axis legend is always placed on top of plot lines and markers.
• By default, its background is fully transparent ( OPAQUE= FALSE), meaning that
underlying lines, markers, and data labels show through the legend.
• Its position can be controlled with the AUTOALIGN= option, or with the HALIGN=
and VALIGN= options.
Within an overlay-type layout, when an axis legend is placed outside the plot area with
LOCATION=OUTSIDE,
• By default, its background is fully opaque (OPAQUE=TRUE).
• Its position can be controlled with the HALIGN= and VALIGN= options.
When an axis legend is placed within nested layouts, it might be necessary to do one of
the following to obtain the desired legend organization:
• use the ACROSS= option and also set ORDER= ROWMAJOR
• use the DOWN= option and also set ORDER=COLUMNMAJOR

Example: AXISLEGEND Statement

The following graph was generated by the “Example Program” on page 1098 . The
LAYOUT OVERLAY statement’s XAXISOPTS= option uses NAME= to assign a name
to the X axis so that it can be referenced in an AXISLEGEND statement. The LAYOUT
OVERLAY statement’s DISCRETEOPTS= option specifies
TICKVALUEFITPOLICY=EXTRACT, which implements the axis legend if a collision
occurs in the tick-value display for the X axis. The AXISLEGEND statement references
the name that was assigned to the X axis.
1098 Chapter 9 • Legend Statements

Example Program
proc template;
define statgraph axislegend ;
begingraph;
entrytitle "Mileage for Vehicles Made by Dodge";
layout overlay / xaxisopts=(name="xaxis"
discreteopts=(tickvaluefitpolicy=extract)) ;
barchart category=model response=mpg_highway / stat=mean ;
axislegend "xaxis";
endlayout;
endgraph;
end;

proc sort data=sashelp.cars out=dodge; by descending mpg_highway;

where make="Dodge";

proc sgrender data=dodge template=axislegend;

run;

CONTINUOUSLEGEND Statement
Creates a legend with a color ramp corresponding to a range of values.

Syntax
CONTINUOUSLEGEND "graph-name" </option(s)>;
CONTINUOUSLEGEND Statement 1099

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.
BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the legend.
INTEGER=TRUE | FALSE
specifies whether only integer tick values are used in the continuous legend.
OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or transparent
(FALSE).
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the legend.
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space that is added outside the legend border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend
perimeter.
VALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend values.

Data range options

VALUECOUNTHINT=positive-integer
recommends a number of values for the continuous legend to use to label the
data range.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the legend is automatically aligned within its parent layout
when nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies the horizontal alignment of the legend within its parent layout when
nested within an overlay-type or region layout.
LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when the
legend is specified within an overlay-type or region layout.
VALIGN=CENTER | TOP | BOTTOM | number
specifies the vertical alignment of the legend within its parent layout when
nested within an overlay-type or region layout.

Scale options
EXTRACTSCALE=TRUE | FALSE
EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC

Text options
TITLE="string"
specifies the title of the legend.
TITLEATTRS=style-element | style-element (text-options) | (text-options)
1100 Chapter 9 • Legend Statements

specifies the color and font attributes of the legend title.

Required Argument
"graph-name"
specifies the plot to be represented by the legend. The plot is identified by the name
that is assigned to it on the plot statement’s NAME= option.

Restriction Unlike the DISCRETELEGEND statement, only a single graph-

name can be specified.

Requirement graph-name must be enclosed in quotation marks.

Optional Arguments
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the legend is automatically aligned within its parent layout when
nested within an overlay-type layout.
NONE
does not automatically align the legend within its parent layout. The legend’s
position is set by the HALIGN= and VALIGN= options.
AUTO
within the parent layout, attempts to center the legend in the area that is farthest
from any surrounding data point markers.
(location-list)
within the parent layout, restricts the legend’s possible locations to those
locations in the specified location-list, and use the location-list position that least
collides with the parent layout’s other graphics features. The location-list is
space-separated and can contain any of these locations: TOPLEFT, TOP,
TOPRIGHT, LEFT, CENTER, RIGHT, BOTTOMLEFT, BOTTOM, and
BOTTOMRIGHT.

Default NONE

Interactions This option has no effect unless LOCATION= INSIDE.

When LOCATION=INSIDE and AUTOALIGN= is not NONE, the

HALIGN= and VALIGN= options are ignored.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type layout.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphLegendBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect for the color to be seen. By default,
OPAQUE=FALSE.
CONTINUOUSLEGEND Statement 1101

BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.

Defaults FALSE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphLegendBackground:FrameBorder style reference starting with

the second maintenance release of SAS 9.4. If attribute FrameBorder is
not defined in the active style, then FALSE is the default value.

Tip The BORDERATTRS= option controls the appearance of the legend

border.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the legend.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

Tip The color of the frame around the color ramp and the color ramp tick
lines is controlled by the GraphAxisLines:contrastColor style attribute.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element

“Line Options” on page 1349 for available line-options.

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it to reduce
the tick value width. The scale can be a named scale or a scientific-notation scale.
The EXTRACTSCALETYPE= option specifies the scale type. The scale used is
appended to the legend title as shown in the following example.
Total Sales (millions)

For long legend titles, if the scale does not fit the available space, then the title is
truncated, and the scale is appended to the truncated title. Ellipses indicate that the
label was truncated as shown in the following example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases where the title does not fit even with truncation, the title is dropped.

Restriction The scale that is extracted by the EXTRACTSCALE= option is

derived from the English locale.

Interactions The scale type is determined by the EXTRACTSCALETYPE= option.

If the axis label is not displayed, then the EXTRACTSCALE=TRUE

option is ignored.

See “boolean ” on page 1339 for other Boolean values that you can use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.
1102 Chapter 9 • Legend Statements

DEFAULT
extracts a named scale. A named scale can be millions, billions, or trillions for
values of 999 trillion or less, or a multiple of 10 (denoted as 10^n) for values
over 999 trillion. For small fractional tick values, the scale factor is set to ensure
that the absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a multiple of
1/10 (10^–n) for values less than 1 trillionth.
SCIENTIFIC
extracts a scientific-notation scale. A scientific-notation scale is a multiple of 10
expressed as 10^n for values greater than 1, or a multiple of 1/10 expressed as
10^–n for values less than 1.
HALIGN=CENTER | LEFT | RIGHT | number
specifies the horizontal alignment of the legend within its parent layout when nested
within an overlay-type or region layout.
number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s width, where 0 is all
the way to the left and 1 is all the way to the right.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. HALIGN=0

corresponds to the containing layout's left edge plus half the legend
width. Similarly, HALIGN=1 corresponds to the containing layout's
right edge minus half the legend width.

Defaults If LOCATION=OUTSIDE, then the default is RIGHT.

If LOCATION=INSIDE, then the default is CENTER.

Restriction This option is available only when this statement is nested within an
overlay-type or region layout.

Interactions If LOCATION= OUTSIDE, then the HALIGN= and VALIGN=

options cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the

AUTOALIGN= option is not NONE, and the parent layout is an
overlay-type layout.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type or region layout.

INTEGER=TRUE | FALSE
specifies whether only integer tick values are used in the continuous legend.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default FALSE
CONTINUOUSLEGEND Statement 1103

Restriction This option applies only to smooth color gradients. For leveled
gradients such as those used with contour plots of type FILL,
LINEFILL or LABELEDLINEFILL, this option is ignored. See
CONTOURTYPE=.

Interaction INTEGER=TRUE is ignored when there are no integer values between

the minimum and maximum legend values.

LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when the legend
is specified within an overlay-type or region layout.

Default OUTSIDE

Restriction This option has effect only when the legend statement appears within
a 2-D overlay-type layout and there is at least one stand-alone plot
statement with XY axes that is referenced by the legend.

Interactions The actual position is determined by the settings for the LOCATION=,
AUTOALIGN= , HALIGN= , and VALIGN= options.

If this option is set to OUTSIDE, then the HALIGN= and VALIGN=

options must specify a keyword (LEFT, RIGHT, or CENTER). The
number setting for the alignment is invalid when the legend is
positioned outside of the plot area.

Within an overlay-type layout, if the ORIENT= option is not set, then

the orientation changes depending on the actual position. If
LOCATION=OUTSIDE and the legend is right or left of the plot, then
the orientation is vertical. If LOCATION=OUTSIDE and the legend is
above or below the plot, then the orientation is horizontal.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type or region layout..

OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the legend.

Default VERTICAL

Restriction In order to use ORIENT=HORIZONTAL when the continuous legend

is in an OVERLAY or REGION layout, you must place the legend
inside the graph area (LOCATION=INSIDE).

Tip To orient the legend horizontally outside of the graph area in an

OVERLAY or REGION layout, use LOCATION=OUTSIDE and
VALIGN=BOTTOM instead.
1104 Chapter 9 • Legend Statements

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space that is added outside the legend border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the legend border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend perimeter.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the legend perimeter.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.
CONTINUOUSLEGEND Statement 1105

Tip Use pad-options to create non-uniform padding.

Default Padding for all sides is 0.

Note The default units for dimension are pixels.

See “dimension” on page 1340

TITLE="string"
specifies the title of the legend. The title is placed below the legend body.

Default No title

Requirement string must be enclosed in quotation marks.

Interaction When the title is placed to the left of the legend, if

TITLEBORDER=TRUE is in effect, no separator is displayed
between the title and the legend. If TITLEBORDER=FALSE is in
effect in that case, a separator is displayed.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend title.

Default The GraphLabelText style element.

Interactions For this option to have any effect, the TITLE= option must also be
specified.

If one or more text options are specified and they do not include all the
font properties (color, family, size, weight, style), then non-specified
properties will be derived from the GraphLabelText style element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

VALIGN=CENTER | TOP | BOTTOM | number

specifies the vertical alignment of the legend within its parent layout when nested
within an overlay-type or region layout.
number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s height, where 0 is on
the bottom and 1 is on the top.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. Zero corresponds to

the containing layout's bottom edge plus half the legend height.
1106 Chapter 9 • Legend Statements

Similarly, one corresponds to the containing layout's top edge minus

half the legend height.

Default CENTER

Restriction This option is available only when this statement is nested within an
overlay-type or region layout. It is ignored if the parent layout is not
an overlay-type or region layout.

Interactions If LOCATION= OUTSIDE, then the VALIGN= and HALIGN=

options cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the

AUTOALIGN= option is not NONE, and the parent layout is an
overlay-type layout.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type or region layout.

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend values.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all the
font properties (color, family, size, weight, style), then non-specified
properties will be derived from the GraphLabelText style element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

VALUECOUNTHINT=positive-integer
recommends a number of values for the continuous legend to use to label the data
range.

Default 6

Restriction The associated plot must be displayed with smooth gradient for this
option to have any effect. For example, in a contour plot,
CONTOURTYPE must be set to GRADIENT or LINEGRADIENT.

Requirement positive-integer must be greater than zero.

Note The legend uses even intervals to label the range.

Details
A continuous legend consists of a color ramp and a numeric scale indicating color
values.
CONTINUOUSLEGEND Statement 1107

In the following figure, the continuous legend references a contour plot with a fixed
number of levels. The color ramp and legend values automatically reflect these discrete
levels. The legend option VALUECOUNTHINT= has no effect.

In this next figure, the continuous legend references a contour plot with a continuous
gradient. The number of legend values displayed is automatically determined by the
legend, or decided by the contributing plot, such as the CONTOURPLOTPARM with
CONTOURTYPE=FILL. For plots with a continuous gradient, you can use the legend
option VALUECOUNTHINT= to control how many legend values appear. (This option
is ignored if there is no gradient.)

Within an overlay-type layout, when a continuous legend is placed inside the plot area
with LOCATION= INSIDE,
• It is always placed on top of plot lines and markers.
• By default, its background is fully transparent ( OPAQUE= FALSE), meaning that
underlying lines, markers, and data labels show through the legend.
1108 Chapter 9 • Legend Statements

• Its position can be controlled with the AUTOALIGN= option, or with the HALIGN=
and VALIGN= options.
Within an overlay-type layout, when a continuous legend is placed outside the plot area
with LOCATION=OUTSIDE,
• By default, its background is fully opaque (OPAQUE=TRUE).
• Its position can be controlled with the HALIGN= and VALIGN= options.

Example: CONTINUOUSLEGEND Statement

The following graph was generated by the “Example Program” on page 1108:

Example Program
proc template;
define statgraph continuouslegend;
begingraph;
entrytitle "Height and Weight Distribution";
layout overlay;
scatterplot x=height y=weight /
markercolorgradient=density
markerattrs=(symbol=squarefilled size=6px)
name="scatter";
continuouslegend "scatter" / orient=vertical
location=outside valign=center halign=right
valuecounthint=10 title="Density";
endlayout;
endgraph;
DISCRETELEGEND and MERGEDLEGEND Statements 1109

end;
run;
ods graphics / reset width=475px;
proc sgrender data=sashelp.gridded(where=(count>0))
template=continuouslegend;
run;

DISCRETELEGEND and MERGEDLEGEND Statements

Creates a legend with entries that refer to plots, or group values, or both legend items.
Restrictions: The MERGEDLEGEND statement can be used for grouped plots only.
The MERGEDLEGEND statement supports only line and marker overlays.
Notes: Often the data in a plot is classified by a group variable. Or, multiple columns of data
are plotted in the same graph. These groups or columns are represented in the
graph by different color or line patterns or marker symbols. In these cases, a
DISCRETELEGEND can be added to the graph to help decode the data. The
MERGEDLEGEND statement can be used to consolidate legend entries when the
graph displays grouped data for two plots. The MERGEDLEGEND statement must
specify exactly two names that reference the source for the legend entry values.
Prior to the third maintenance release of SAS 9.4, when a discrete attribute map is
used for group values in a plot that contributes to a discrete legend and attributes are
overridden in the plot statement, the attributes of some plot features and their
corresponding legend items might not match. Starting with the third maintenance
release of SAS 9.4, the attributes of the legend items always match the attributes of
the corresponding plot features.
See: “LEGENDITEM Statement” on page 1126

Syntax
DISCRETELEGEND "graph-name-1" | "legend-item-name-1" | "discrete-attr-name-1"
<"graph-name-2" | "legend-item-name-2" | "discrete-attr-name-2"…> </option(s)>;
MERGEDLEGEND "graph-name" | "legend-item-name" | "discrete-attr-name"
"graph-name" | "legend-item-name" | "discrete-attr-name"
</option(s)>;

Summary of Optional Arguments

Appearance options
ACROSS=positive-integer
specifies the number of legend entries that are placed horizontally before the
next row begins.
AUTOITEMSIZE=TRUE | FALSE
specifies that all markers, lines, and filled symbols in the legend are sized in
proportion to the font size used for the legend entry labels.
BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.
BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
1110 Chapter 9 • Legend Statements

specifies the attributes of the border line around the legend.

DISPLAYCLIPPED=TRUE | FALSE
specifies whether the legend is displayed when any portion of the legend
cannot be rendered entirely.
DOWN=positive-integer
specifies the number of legend entries that are placed vertically before the
next column begins.
FILLITEMOUTLINE=AUTO | ON
specifies whether the fill swatches are outlined only when enabled by the
contributing statements or are always outlined.
ITEMSIZE=(size-options)
specifies the size of specific types of items that are in a discrete or merged
legend.
OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or transparent
(FALSE).
ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether legend entries are organized into rows or into columns.
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space that is added outside the legend border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend
perimeter.
SORTBY=LABEL | TEXT
specifies whether text legend items are sorted by label or by text.
SORTORDER=AUTO | ASCENDINGFORMATTED |
DESCENDINGFORMATTED
specifies the sort order to use for the legend entry labels.
TITLEBORDER=TRUE | FALSE
specifies a border around the legend title that separates it from the legend
body.
VALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend values.

Content options
ADDITIONALNAMES=("graph-name" | "legend-item-name" | "discrete-attr-name"
<"graph-name" | "legend-item-name" | "discrete-attr-name">…)
specifies additional legend items that are to be added to the two items that are
required in the MERGEDLEGEND statement.
EXCLUDE=(item-names)
specifies a list of legend entries to exclude from the display.
TYPE=ALL | FILL | FILLCOLOR | LINE | LINECOLOR | LINEPATTERN |
MARKER | MARKERCOLOR | MARKERSYMBOL | TEXT
specifies which visual attributes to display for legend entries in the legend.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the legend is automatically aligned within its parent layout
when nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies the horizontal alignment of the legend within its parent layout when
nested within an overlay-type or region layout.
DISCRETELEGEND and MERGEDLEGEND Statements 1111

LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when the
legend is specified within an overlay-type or region layout.
VALIGN=CENTER | TOP | BOTTOM | number
specifies the vertical alignment of the legend within its parent layout when
nested within an overlay-type or region layout.

Text options
TITLE="string"
specifies the title of the legend.
TITLEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend title.

Required Arguments
"graph-name"
references one or more unique names that are specified by the NAME= option of a
plot statement. These names control which plots contribute to the legend, and the
order in which the legend entries occur. If a contributing plot uses a GROUP=
option, then there is a legend entry for each group value that is present in the data.

Requirements Each graph-name must be enclosed in quotation marks.

Each plot that is referenced must be able to be identified by the

legend entries. For example, a filled CONTOURPLOTPARM plot
cannot be specified in the DISCRETELEGEND or
MERGEDLEGEND statement because it requires a continuous
legend.

"legend-item-name"
references one or more unique values specified by the NAME= option of a
LEGENDITEM statement. Each legend-item-name must be enclosed in quotation
marks.
"discrete-attr-name"
references one or more unique values that are specified by the NAME= option in a
DISCRETEATTRMAP statement. The discrete attribute map that the name
references contributes all of its value statements as legend items, regardless of
whether they match the data.

Optional Arguments
ACROSS=positive-integer
specifies the number of legend entries that are placed horizontally before the next
row begins. A legend entry typically consists of two parts, such as a marker symbol
and an associated value.

Default The entries are placed to best fit the available area. This “best fit”
approach works only when the legend is nested in the template’s
outermost layout.

Interactions This option is ignored if ORDER= COLUMNMAJOR

This option is ignored when the DISCRETELEGEND statement is in

a LAYOUT GLOBALLEGEND block and the TYPE=ROW option is
in effect for the layout.
1112 Chapter 9 • Legend Statements

ADDITIONALNAMES=("graph-name" | "legend-item-name" | "discrete-attr-

name" <"graph-name" | "legend-item-name" | "discrete-attr-name">…)
specifies additional legend items that are to be added to the two items that are
required in the MERGEDLEGEND statement.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
The items from the sources that are specified in this option are not included in the
merging of the legend items. They are appended to the legend after the items from
the sources that are specified in the required arguments are merged.

Restriction This option applies to the MERGEDLEGEND statement only.

Requirements The list of names must be enclosed in parentheses.

Each name must be enclosed in quotation marks and separated from

adjacent values by a blank space.

See “"graph-name"” on page 1111

“"legend-item-name"” on page 1111

“"discrete-attr-name"” on page 1111

AUTOALIGN=NONE | AUTO | (location-list)

specifies whether the legend is automatically aligned within its parent layout when
nested within an overlay-type layout.
NONE
does not automatically align the legend within its parent layout. The legend’s
position is set by the HALIGN= and VALIGN= options.
AUTO
within the parent layout, attempts to center the legend in the area that is farthest
from any surrounding data point markers.
(location-list)
within the parent layout, restricts the legend’s possible locations to those
locations in the specified location-list, and use the location-list position that least
collides with the parent layout’s other graphics features. The location-list is
space-separated and can contain any of these locations: TOPLEFT, TOP,
TOPRIGHT, LEFT, CENTER, RIGHT, BOTTOMLEFT, BOTTOM, and
BOTTOMRIGHT.

Default NONE

Restriction AUTO is available only when the parent layout contains a scatter plot.
Otherwise, it is ignored.

Interactions This option has no effect unless LOCATION= INSIDE.

When LOCATION=INSIDE and AUTOALIGN= is not NONE, the

HALIGN= and VALIGN= options are ignored.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type layout.
DISCRETELEGEND and MERGEDLEGEND Statements 1113

AUTOITEMSIZE=TRUE | FALSE
specifies that all markers, lines, and filled symbols in the legend are sized in
proportion to the font size used for the legend entry labels. These proportional sizes
take effect regardless of the size reported by the plot or LEGENDITEM. The line
segments are drawn as deemed appropriate by the legend, regardless of the line
thickness reported by the plot.

Default FALSE

Interaction When set to TRUE, this setting considers the font size in effect from
the VALUEATTRS= option.

See “boolean ” on page 1339 for other Boolean values that you can use.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphLegendBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect for the color to be seen. By default,
OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.

Defaults TRUE in the first maintenance release of SAS 9.4 and earlier releases.

The GraphLegendBackground:FrameBorder style reference starting with

the second maintenance release of SAS 9.4. If attribute FrameBorder is
not defined in the active style, then TRUE is the default value.

Tip The BORDERATTRS= option controls the appearance of the legend

border.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the legend.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element

“Line Options” on page 1349 for available line-options.

DISPLAYCLIPPED=TRUE | FALSE
specifies whether the legend is displayed when any portion of the legend cannot be
rendered entirely. Based on the legend contents and placement, or when the graph
size is reduced, parts of the legend (title, legend symbol, or legend value) might be
clipped (truncated). When clipping occurs and this option is set to FALSE, the entire
legend is removed from the graph. The space for that legend is then reclaimed by the
1114 Chapter 9 • Legend Statements

remainder of the graph. When this option is set to TRUE, the legend always appears,
even if some parts of it have been clipped.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

DOWN=positive-integer
specifies the number of legend entries that are placed vertically before the next
column begins. A legend entry typically consists of two parts, such as a marker
symbol and an associated value.

Default The entries are placed to best fit the available area. This “best fit”
approach works only when the legend is nested in the template’s
outermost layout.

Restriction This option is ignored if ORDER= ROWMAJOR

Interaction This option is ignored when the DISCRETELEGEND statement is in a

LAYOUT GLOBALLEGEND block.

EXCLUDE=(item-names)
specifies a list of legend entries to exclude from the display.

Default No items are excluded.

Requirement Each item name must be enclosed in quotation marks and separated
from adjacent names by a space.

Note When the specified names are compared with the legend entry names,
leading blanks are honored and trailing blanks are ignored.

Tip For plots with groups, you can exclude specific group values.

Example The following example excludes items Truck and Wagon from the
legend.
exclude=("Truck" "Wagon")

FILLITEMOUTLINE=AUTO | ON
specifies whether the fill swatches are outlined only when enabled by the
contributing statements or are always outlined.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

AUTO honors the DISPLAY= option settings for the contributing plot or the
FILLDISPLAY= setting of the legend item.
ON always outlines the fill swatches.

Default AUTO

Restriction This option is valid only in the DISCRETELEGEND statement.

Interaction The legend must display fill entries for this option to have any effect.

Note The outlines are always 1 pixel wide with a solid pattern.
DISCRETELEGEND and MERGEDLEGEND Statements 1115

HALIGN=CENTER | LEFT | RIGHT | number

specifies the horizontal alignment of the legend within its parent layout when nested
within an overlay-type or region layout.
number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s width, where 0 is all
the way to the left and 1 is all the way to the right.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. HALIGN=0

corresponds to the containing layout's left edge plus half the legend
width. Similarly, HALIGN=1 corresponds to the containing layout's
right edge minus half the legend width.

Default CENTER

Restriction This option is available only when this statement is nested within an
overlay-type or region layout.

Interactions If LOCATION= OUTSIDE, then the HALIGN= and VALIGN=

options cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the

AUTOALIGN= option is not NONE, and the parent layout is an
overlay-type layout.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type or region layout.

ITEMSIZE=(size-options)
specifies the size of specific types of items that are in a discrete or merged legend.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
The following size options are supported:
FILLASPECTRATIO=AUTO | GOLDEN | positive-number
specifies the aspect ratio for the fill swatches.
Note: This feature applies to the third maintenance release of SAS 9.4 and to
later releases.
AUTO
uses an equal width and height for color swatches with solid color fills, or
uses the golden ratio for swatches with pattern fills.
GOLDEN
specifies the golden ratio of 1.618 (width = 1.618 * height) for both solid
color and pattern fill swatches.
positive-number
specifies a custom aspect ratio.
1116 Chapter 9 • Legend Statements

Default AUTO

Restrictions This option is valid only in the DISCRETELEGEND statement.

This option does not apply to bubble plot fill color swatches (filled
circle).

Interactions The legend must display fill entries for this option to have any
effect.

This option is ignored when AUTOITEMSIZE=FALSE.

When FILLASPECTRATIO= is specified but neither

FILLHEIGHT= nor HEIGHTSCALE= is used, the height for
color swatches with a solid fill and color swatches with a pattern
fill differ.

Note FILLHEIGHT=, HEIGHTSCALE=, and FILLASPECTRATIO=

also apply to fill-pattern color swatches with no outline.

Tip Use FILLHEIGHT= or HEIGHTSCALE= to specify the height.

FILLHEIGHT=AUTO | BIG | dimension | style-reference

specifies the height of the fill swatches.
Note: This feature applies to the third maintenance release of SAS 9.4 and to
later releases.
AUTO
enables the SAS system to determine the fill color swatch height.
BIG
specifies sizes that were developed as defaults for SAS Visual Analytics,
which results in color swatches that are larger than the SAS system defaults.
dimension
specifies a custom height for the fill swatches.

See “dimension” on page 1340

style-reference
specifies a style attribute that controls the height of the fill swatches. The
style reference must provide a valid dimensional value.

See “style-reference ” on page 1342

Default AUTO

Restriction This option is valid only in the DISCRETELEGEND statement.

Interactions The legend must display fill entries for this option to have any
effect.

This option is ignored when AUTOITEMSIZE=FALSE.

Tip Use FILLASPECTRATIO= to specify the width.

DISCRETELEGEND and MERGEDLEGEND Statements 1117

HEIGHTSCALE=positive-number
specifies a scale factor that is to be applied to the fill swatch height. Values
greater than 1 increase the height of the fill swatches, and values less than 1
reduce the height.
Note: This feature applies to the third maintenance release of SAS 9.4 and to
later releases.

Default 1.5

Restriction This option is valid only in the DISCRETELEGEND statement.

Interactions The legend must display fill entries for this option to have any
effect.

This option is ignored when AUTOITEMSIZE=FALSE.

Tips Use FILLHEIGHT= to change the base height.

Use FILLASPECTRATIO= to specify the width.

LINELENGTH=dimension
specifies the length of the line glyph for line entries in the legend.

Default Determined by the system

Interaction The legend must display line entries for this option to have any
effect.

See “dimension” on page 1340

LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when the legend
is specified within an overlay-type or region layout.

Default OUTSIDE

Restriction This option has effect only when the legend statement appears within
an overlay-type or region layout and at least one stand-alone plot
statement is referenced by the parent layout.

Interactions The actual position is determined by the settings for the LOCATION=,
AUTOALIGN= , HALIGN= , and VALIGN= options.

If this option is set to OUTSIDE, then the HALIGN= and VALIGN=

options must specify a keyword (LEFT, RIGHT, or CENTER). The
number setting for the alignment is invalid when the legend is
positioned outside of the plot area.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type or region layout..

OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or transparent (FALSE).

Default FALSE
1118 Chapter 9 • Legend Statements

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether legend entries are organized into rows or into columns.

Default ROWMAJOR

Interaction If ORDER=ROWMAJOR, then use the ACROSS= option to limit the

number of entries in a row. If ORDER=COLUMNMAJOR, then use
the DOWN= option to limit the number of entries in a column.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space that is added outside the legend border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the legend border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Notes The default units for dimension are pixels.

Starting with the first maintenance release of SAS 9.4, the default padding
between the discrete legend and the plot area (including the axes) is
increased to 10 pixels, depending on the context. If the new default
padding is not desirable, then use the OUTERPAD= option to adjust it.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend perimeter.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the legend perimeter.
DISCRETELEGEND and MERGEDLEGEND Statements 1119

(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default Padding for all sides is 0.

Note The default units for dimension are pixels.

See “dimension” on page 1340

SORTBY=LABEL | TEXT
specifies whether text legend items are sorted by label or by text.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.

Default LABEL

Restriction This option is valid only in the DISCRETELEGEND statement.

Interaction This option has an effect only when sorting is performed,

TYPE=TEXT is specified in this DISCRETELEGEND statement, and
the legend items are defined in either a LEGENDTEXTITEM
statement or in a LEGENDITEM statement that specifies
TYPE=TEXT.

SORTORDER=AUTO | ASCENDINGFORMATTED |
DESCENDINGFORMATTED
specifies the sort order to use for the legend entry labels.

Default AUTO. Groups of legend entries are presented in the order in which
they are listed in the legend statement. The internal ordering of the
entries is derived from the constituent plot-statement options.

Interactions This option overrides the order that is set by any constituent plot
statement’s GROUPORDER= option.

If this option is set to ASCENDINGFORMATTED or

DESCENDINGFORMATTED, then the entries from separate plots,
and separate legend items are combined and ordered as a single list.
1120 Chapter 9 • Legend Statements

Note The ASCENDINGFORMATTED and DESCENDINGFORMATTED

settings perform a linguistic sort on the group items and have the same
effect as sorting the input data. However, the data is not changed.

TITLE="string"
specifies the title of the legend. The title is placed to the left of the legend body,
except in the following cases:
• The legend contains two or more rows of items.
• The legend is in a nested layout.
• The legend is in an OVERLAYEQUATED layout.
• The legend title length exceeds the space that is available on the left side of the
legend.
In those cases, the title is placed above the legend body.

Default No title

Requirement string must be enclosed in quotation marks.

Interaction When the title is placed to the left of the legend, if

TITLEBORDER=TRUE is in effect, no separator is displayed
between the title and the legend. If TITLEBORDER=FALSE is in
effect in that case, a separator is displayed.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend title.

Default The GraphLabelText style element.

Interactions For this option to have any effect, the TITLE= option must also be
specified.

If one or more text options are specified and they do not include all the
font properties (color, family, size, weight, style), then non-specified
properties will be derived from the GraphLabelText style element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Text Options” on page 1351 for available text-options.

TITLEBORDER=TRUE | FALSE
specifies a border around the legend title that separates it from the legend body.

Default FALSE

Interactions Prior to the third maintenance release of SAS 9.4, when

LOCATION=INSIDE, BORDER=TRUE and the TITLE= option must
also be in effect for this option to have any effect. When
LOCATION=OUTSIDE, the TITLE= option must also be in effect.
The BORDER= option has no effect on the title border in that case.
DISCRETELEGEND and MERGEDLEGEND Statements 1121

Starting with the third maintenance release of SAS 9.4,

BORDER=TRUE and the TITLE= option must also be in effect for
this option to have any effect. The LOCATION= option has no effect
on the title border in that case.

Tip The line attributes of the title border are set by the BORDERATTRS=
options.

See “boolean ” on page 1339 for other Boolean values that you can use.

TYPE=ALL | FILL | FILLCOLOR | LINE | LINECOLOR | LINEPATTERN |

MARKER | MARKERCOLOR | MARKERSYMBOL | TEXT
specifies which visual attributes to display for legend entries in the legend.
Note: TEXT is valid starting with the third maintenance release of SAS 9.4.
The TYPE= option can be used as a filter. If a statement contributing to the legend
does not have any visual attributes that match the TYPE specified, then the legend
does not display any entries from that statement.
Some keywords can be used to create specialized legends that display a single visual
attribute. For example, keywords FILLCOLOR or MARKERSYMBOL result in the
display of a single attribute. Other keywords (for example, FILL, LINE, or
MARKER) result in legends that display a set of visual attributes. For example,
keyword LINE results in the display of both line color and line pattern for legend
entries that include lines in their display.
If this option is set to LINEPATTERN or MARKERSYMBOL, then a filled symbol
is drawn using the same text color as the color used for the legend entry labels. The
symbol is sized automatically, as if the AUTOITEMSIZE= option is set to TRUE.
For keywords FILLCOLOR, LINECOLOR, and MARKERCOLOR, the filled
symbols are drawn as outlined color swatches. The outline is 1 pixel wide, and its
color is controlled by the CONTRASTCOLOR attribute of the GraphOutlines style
element.

Default ALL

Restrictions This option is valid only in the DISCRETELEGEND statement.

A LEGENDITEM statement can be referenced only from a discrete

legend of the same attribute type or of an overlapping attribute type.
For legends that display multiple visual attributes (use both colors and
marker symbols, for example), the default visual properties are
derived from the GraphDataDefault style-element.

Starting with the third maintenance release of SAS 9.4, the

LEGENDTEXTITEMS statement can be referenced only from a
discrete legend of type TEXT or ALL.

VALIGN=CENTER | TOP | BOTTOM | number

specifies the vertical alignment of the legend within its parent layout when nested
within an overlay-type or region layout.
number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s height, where 0 is on
the bottom and 1 is on the top.
1122 Chapter 9 • Legend Statements

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. Zero corresponds to

the containing layout's bottom edge plus half the legend height.
Similarly, one corresponds to the containing layout's top edge minus
half the legend height.

Default CENTER

Restriction This option is available only when this statement is nested within an
overlay-type or region layout. It is ignored if the parent layout is not
an overlay-type or region layout.

Interactions If LOCATION= OUTSIDE, then the VALIGN= and HALIGN=

options cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the

AUTOALIGN= option is not NONE, and the parent layout is an
overlay-type layout.

See the LAYOUT OVERLAY “LAYOUT OVERLAY Statement” on page

142 for more information about how child positions are determined in
an overlay-type or region layout.

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend values.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all the
font properties (color, family, size, weight, style), then non-specified
properties will be derived from the GraphLabelText style element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

Details
A discrete legend consists of one or more units called legend entries. Each legend entry
consists of a legend symbol and a legend value. The legend symbol is typically a marker,
line, or filled symbol that represents a specific area in the plot. The legend value is
descriptive text that is derived from group values, or that is assigned by the plot’s
LEGENDLABEL= option or by the LEGENDITEM statement’s LABEL= option. To
specify a discrete legend, you can use the DISCRETELEGEND statement. To
consolidate legend entries for common grouped values that are represented by two
separate plots, you can use the MERGEDLEGEND statement.
For grouped plots, a discrete legend represents all of the values that are present in the
data. To ensure that legend entries are displayed for group values, regardless of whether
DISCRETELEGEND and MERGEDLEGEND Statements 1123

those values are present in the data, you can use a discrete attribute map as described in
“Displaying Legend Entries for Group Values That Are Not in the Data” on page 1293.
In the DISCRETELEGEND statement, you can specify one or more names that
reference the source for the legend entry values. You can use the TYPE= option to
control the visual attributes for the legend display.
For legend items that represent fills or colors without a specific shape, a filled symbol
with a one-pixel-wide, solid outline is used to represent the data values. If the feature
being represented by the legend item has an outline, then the default outline color is
derived from the feature’s outline color. If the feature being represented is an outline
only or is a filled outlined marker, then the outline thickness is derived from the plot. If
the feature does not have an outline, then the default outline color for its corresponding
legend entry is derived from the GraphOutline style element.
You can use the MERGEDLEGEND statement to consolidate lines and marker symbols
from discrete legend entries when a graph displays grouped values for exactly two plots.
With a MERGEDLEGEND statement, the legend values from the group variables in two
plots are compared. For each common value, the corresponding legend lines and marker
symbols are combined, and only one legend entry is created for each matching set of
group values. You can use the MERGEDLEGEND statement only for grouped plots.
You must specify two names that reference the source for the legend entry values.
When a discrete legend is placed inside the plot area of an overlay type layout and
LOCATION= INSIDE, the following conditions apply:
• The discrete legend is always placed on top of plot lines and markers.
• By default, the background of the discrete legend is fully transparent ( OPAQUE=
FALSE). Underlying lines, markers, and data labels show through the legend.
• You can control the position of the discrete legend with the AUTOALIGN= option,
or with the HALIGN= and VALIGN= options. (The AUTOALIGN= option is not
available in a LAYOUT OVERLAY3D statement.)
When a discrete legend is placed outside the plot area of an overlay type layout and
LOCATION=OUTSIDE, the following conditions apply:
• By default, the background of the discrete legend is fully opaque
(OPAQUE=TRUE).
• You can control the position of the discrete legend with the HALIGN= and
VALIGN= options.
When a discrete legend is placed within nested layouts, you might need to do one of the
following to obtain the legend organization that you want:
• use the ACROSS= option, and also set ORDER= ROWMAJOR
• use the DOWN= option, and also set ORDER=COLUMNMAJOR
A legend might be dropped if the total legend area in the graph exceeds the percentage
that is set by the MAXLEGENDAREA= option in an ODS GRAPHICS ON statement
that is in effect for the output destination. A legend might also be dropped if
DISPLAYCLIPPED= FALSE and the full legend cannot be displayed.
1124 Chapter 9 • Legend Statements

Examples

Example 1: DISCRETELEGEND Statement

The following graph using the DISCRETELEGEND statement was generated by
“Example Program” . It displays two discrete legends, one that shows the confidence
limits for two ellipses and a second that shows the values for a GROUP= variable:

Example Program
proc template;
define statgraph discretelegend;
begingraph;
entrytitle "Prediction Ellipses";
layout overlayequated / equatetype=equate;
scatterplot x=petallength y=petalwidth /
group=species name="s";
ellipse x=petallength y=petalwidth /
type=predicted alpha=0.2
name="p80" legendlabel="80%"
outlineattrs=graphconfidence;
ellipse x=petallength y=petalwidth /
type=predicted alpha=0.05
name="p95" legendlabel="95%"
outlineattrs=graphconfidence2;
discretelegend "s" / title="Species:" ;
discretelegend "p80" "p95" / across=1
autoalign=(topleft) location=inside ;
endlayout;
entryfootnote halign=left "Fisher's Iris Data";
endgraph;
end;
Example 2: MERGEDLEGEND Statement 1125

proc sgrender data=sashelp.iris template=discretelegend;

run;

Example 2: MERGEDLEGEND Statement

The following graph using the MERGEDLEGEND statement was generated by
“Example Program” . In the template definition, a grouped scatter plot is overlaid with a
series plot for each group, and the two plots are referenced by a single merged legend:

Example Program
proc sort data=sashelp.class
out =class;
by sex;
run;

proc template;
define statgraph mergedLegend;
begingraph;
entrytitle "Linear Regression By Gender";
layout overlay;
scatterplot x=height y=weight / group=sex name="scat";
regressionplot x=height y=weight/ group=sex name="reg";
mergedlegend "scat" "reg" / border=true;
endlayout;
endgraph;
end;

proc sgrender data=class template=mergedLegend;

run;
1126 Chapter 9 • Legend Statements

LEGENDITEM Statement
Creates the definition for a legend item that can be included in a discrete legend.
Restriction: The LEGENDITEM statement is used with the DISCRETELEGEND and
MERGEDLEGEND statements only.
Requirement: The LEGENDITEM statement must appear in the global definition area of the
template between the BEGINGRAPH statement and the first LAYOUT statement.
Note: The LEGENDITEM statement creates the definition for a legend item that can be
included in a discrete legend.
See: “DISCRETELEGEND and MERGEDLEGEND Statements” on page 1109

Syntax
LEGENDITEM TYPE=type NAME="string" </option(s)>;

Summary of Optional Arguments

Appearance options
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the color of the fill when TYPE= is set to FILL.
FILLDISPLAY=STANDARD | ALL | (FILL | OUTLINE)
specifies whether the fill swatch for this legend item displays fill only or
displays fill and outline.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the line when TYPE= is set to LINE or
MARKERLINE.
MARKERATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the appearance of the marker when TYPE= is set to MARKER or
MARKERLINE.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the color of the outline when TYPE= is set to FILL.
TEXT="string"
specifies the text to display when TYPE= is set to TEXT.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
specifies the font and color attributes of the string that is specified on the
TEXT= option when TYPE= is set to TEXT.

Label options
LABEL="string"
specifies a label to be used for the legend item.
LABELATTRS
specifies the color and font attributes of the legend item’s label.
LEGENDITEM Statement 1127

Required Arguments
TYPE=FILL | MARKER | MARKERLINE | LINE | TEXT
specifies a type for the legend item.
FILL
specifies a fill (displayed as a filled square). The appearance can be managed
with the FILLATTRS= and OUTLINEATTRS= options.
MARKER
specifies a marker. The appearance can be managed with the MARKERATTRS=
option.
MARKERLINE
specifies a marker and a line. The appearance can be managed with the
MARKERATTRS= and LINEATTRS= options.
LINE
specifies a line. The appearance can be managed with the LINEATTRS= option.
TEXT
specifies text that can be displayed in the legend area. The text string is defined
with the TEXT= option, and the text appearance can be managed with the
TEXTATTRS= option.
NAME="string"
assigns a name to the legend item for reference in a DISCRETELEGEND or
MERGEDLEGEND statement.

Restriction The string is case sensitive and must define a unique name within the
template.

Optional Arguments
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the color of the fill when TYPE= is set to FILL.

Default The GraphDataDefault style element.

Interaction The TRANSPARENCY attribute cannot be derived from the style

element, but it can be set with this option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

FILLDISPLAY=STANDARD | ALL | (FILL | OUTLINE)

specifies whether the fill swatch for this legend item displays fill only or displays fill
and outline.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
STANDARD | ALL
displays the fill and outline.
(FILL | OUTLINE)
displays only the fill (FILL) or displays the fill and outline (OUTLINE).

Default STANDARD
1128 Chapter 9 • Legend Statements

Requirement You must specify TYPE=FILL for this option to have any effect.

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline.

Interaction When this option is TRUE, the marker fill is drawn using the
FILLATTRS= specification, and the outline is drawn using the
OUTLINEATTRS= specification.

FALSE
draws the markers using fill or an outline, but not both.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

LABEL="string"
specifies a label to be used for the legend item.

Default No label is displayed

Tip The font and color attributes for the label are specified by the
LABELATTRS= option.

LABELATTRS
specifies the color and font attributes of the legend item’s label.

Default The GraphValueText style element.

Interaction For this option to take effect, the LABEL= option must also be
specified.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the line when TYPE= is set to LINE or MARKERLINE.

Default The GraphDataDefault style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

MARKERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the appearance of the marker when TYPE= is set to MARKER or
MARKERLINE.

Default The GraphDataDefault style element.

LEGENDITEM Statement 1129

Interaction The WEIGHT attribute cannot be derived from the style element, but it
can be set with this option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Marker Options” on page 1350 for available marker-options.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the color of the outline when TYPE= is set to FILL.

Default The GraphDataDefault style element.

Restriction This option uses only the color specification in the style element or line
options. The line pattern and line thickness specifications are ignored.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

TEXT="string"
specifies the text to display when TYPE= is set to TEXT. The font and color
attributes for the text are specified by the TEXTATTRS= option.

Default a blank space

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the font and color attributes of the string that is specified on the TEXT=
option when TYPE= is set to TEXT.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

Details
The LEGENDITEM statement creates a definition for a legend item that can be included
in a legend. The item that you define is independent of the data and enables you to
customize the legend to enhance or replace a standard legend. For example, to display
annotation text within the legend area, you can define a LEGENDITEM statement with
TYPE=TEXT and specify the text in the TEXT= option.
As demonstrated in the “Example Program” on page 1130, multiple LEGENDITEM
statements can be used to replace a plot statement’s data-driven legend by defining
custom legend items to display in the legend. This use enables you to communicate
information that is not in the data. For this use, you would define one or more
LEGENDITEM statements to specify legend attributes, and then reference those items in
your legend statement. You must not reference the plot statement itself in the legend
statement. Although no direct connection would exist between the plot data and the
legend, you could communicate the connection by managing the visual attributes in both
the plot and the legend.
The LEGENDITEM statement must be located within the BEGINGRAPH block but
outside of the outermost layout block. You can use multiple LEGENDITEM statements
1130 Chapter 9 • Legend Statements

to define multiple definitions. In that case, each definition specifies a single legend entry
and each item name must be referenced in the legend statement.
Note: A single legend statement can reference multiple item names and also multiple
plot names.
When specifying attribute options for a particular LEGENDITEM statement, options
that do not apply to the specified TYPE= value are ignored. For example, the
MARKERATTRS= option sets the attributes for a marker and is useful if you set
TYPE=MARKER or TYPE=MARKERLINE. However, if TYPE= is set to a value that
does not display a marker symbol, then the MARKERATTRS= option is ignored.

Example: LEGENDITEM Statement

The following graph was generated by the “Example Program” on page 1130. The
example specifies three LEGENDITEM statements to define graphical properties for
two marker symbols and a text string. The NAME= option in each LEGENDITEM
statement assigns a name to the definition. Those names are referenced in a
DISCRETELEGEND statement so that the definitions are displayed in the graph legend.
To correlate the legend with the scatter plot, the example creates an attribute map that
matches values M and F to the same graphical properties that are specified in the
LEGENDITEM statements. That attribute map is referenced in the scatter plot.

Example Program
proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Team Tryouts: Height and Weight by Sex";
discreteattrmap name="symbols" / ignorecase=true trimleading=true;
LEGENDTEXTITEMS Statement 1131

value "m" / markerattrs=(color=blue symbol=diamondfilled);

value "f" / markerattrs=(color=red symbol=circlefilled);
enddiscreteattrmap;
discreteattrvar attrvar=groupmarkers var=sex attrmap="symbols";
legendItem type=marker name="m_marker" /
markerattrs=(color=blue symbol=diamondfilled)
label="70% of boys present" ;
legendItem type=marker name="f_marker" /
markerattrs=(color=red symbol=circlefilled)
label="75% of girls present" ;
legendItem type=text name="status" /
text=" (5 health forms missing)" ;
layout overlay;
scatterplot x=height y=weight / group=groupmarkers;
discretelegend "m_marker" "f_marker" "status" /
autoitemsize=true;
endlayout;
endgraph;
end;
proc sgrender data=sashelp.class template=scatterplot;
run;

LEGENDTEXTITEMS Statement
Creates the definition for data-driven text items that can be included in a discrete legend.
Restrictions: The LEGENDTEXTITEMS statement is used only with the DISCRETELEGEND
statement.
Grouping is not supported.
The maximum number of items that the LEGENDTEXTITEMS statement can
contribute to the legend is 100. If this limit is exceeded, the LEGENDTEXTITEMS
statement is dropped. If no other statements contribute to the legend in that case,
the legend is not drawn.
Requirements: The LEGENDTEXTITEMS statement must appear in the global definition area of the
template between the BEGINGRAPH statement and the first LAYOUT statement.
The TYPE= option in the legend statement that references this statement must be
set to ALL or TEXT.
Note: This feature applies to the third maintenance release of SAS 9.4 and to later
releases.
See: “DISCRETELEGEND and MERGEDLEGEND Statements” on page 1109

Syntax
LEGENDTEXTITEMS NAME="string" TEXT=column </option(s)>;

Required Arguments
NAME="string"
assigns a name to the legend items for reference in a DISCRETELEGEND or
MERGEDLEGEND statement.
1132 Chapter 9 • Legend Statements

TEXT=column
specifies the column in the plot data set that contains the text items.

Notes One item is added for each observation. Grouping is not supported.

The TEXT column should not contain missing values. A missing TEXT
column value is treated as if no text is specified. If an observation contains a
missing TEXT column value and a valid LABEL column value, only the
label value is added to the legend for that observation. If both values are
missing, nothing is added to the legend.

Optional Arguments
LABEL=column
specifies the column that contains the labels for the legend items.

Default No labels are displayed for the items

Note Each observation that has a LABEL column value should have a
corresponding TEXT column value. If an observation contains a valid
LABEL column value and a missing TEXT column value, only the label
value is added to the legend for that observation.

Tip The font and color attributes for the label are specified by the
LABELATTRS= option.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend item labels.

Default The GraphValueText style element.

Note Space is reserved in the legend for the height of the legend label text
regardless of whether the LABEL= option is specified.

Tip If you are not using the LABEL= option, specify

LABELATTRS=(SIZE=0pt) to reclaim the space that is reserved for the
label text height.

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element value.

“Text Options” on page 1351 for available text-options.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the font and color attributes of the text values that is specified on the
TEXT= option.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax for
using a style-element value.

“Text Options” on page 1351 for available text-options.

Example: LEGENDTEXTITEMS Statement 1133

Details
The LEGENDTEXTITEMS statement creates a definition for data-driven text legend
items that can be included in a discrete legend. The items that you define in the data are
independent of the plot and enable you to customize the legend in order to enhance or
replace a standard legend. The text items are stored in the plot data set in the column that
is specified by the TEXT= option. The LEGENDTEXTITEMS statement must be
located within the BEGINGRAPH block but outside of the outermost layout block. You
reference the LEGENDTEXTITEMS statement in your legend statement by the name
specified in the NAME= option. You must not reference the plot statement itself in the
legend statement. To add items from multiple columns, specify one
LEGENDTEXTITEMS statement for each column, and then reference all of the
LEGENDTEXTITEMS statements in your legend statement. Although no direct
connection would exist between the plot data and the legend, you could communicate
the connection by managing the visual attributes in both the plot and the legend.

Example: LEGENDTEXTITEMS Statement

The following graph was generated by the “Example Program” on page 1133. The
graph shows a scatter plot of height and weight by sex for 19 participants in a team
tryout. To reduce clutter in the plot, a numeric ID is used to label the marker for each
individual in the plot. A legend in the right margin of the graph displays the participant
name for each ID.

Example Program
In this example program, a LEGENDTEXTITEMS statement is used to create the legend
of IDs and names in the right margin of the graph. The name IDLEGEND is assigned to
the LEGENDTEXTITEMS statement. The ID column provides the legend text items,
and the Name column provides a name as the label for each text item. Here is the SAS
code.
/* Assign a numeric ID to each name */
1134 Chapter 9 • Legend Statements

data class;
set sashelp.class;
id=_N_;
run;

/* Define the template for the graph */

proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Team Tryouts: Height and Weight by Sex";
legendtextItems name="idlegend" text=id / label=name;
layout lattice / columns=2 columnweights=(85 15);
layout overlay /
xaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot))
yaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot));
scatterplot x=height y=weight / name="scatter"
group=sex datalabel=id;
discretelegend "scatter" / location=inside
autoalign=(bottom bottomright);
endlayout;
layout overlay;
discretelegend "idlegend" /
title="ID/Name" titleattrs=(weight=bold)
valign=top border=false
order=rowmajor across=1;
endlayout;
endlayout;
endgraph;
end;

/* Render the graph */

proc sgrender data=class template=scatterplot;
run;
1135

Part 7

Text Statements

Chapter 10
Managing Text Items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1137

Chapter 11
Text Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1147
1136
1137

Chapter 10
Managing Text Items

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1137
Using Prefix Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1138
Available Prefix Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1138
Managing Horizontal Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1138
Managing Font Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1139
Using Text Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1140
Available Text Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1140
Subscripting and Superscripting Text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1141
Using UNICODE Text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1141
Rules for Unicode and Special Character Specifications . . . . . . . . . . . . . . . . . . . 1142
Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1142
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1142
Lowercase Greek Letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1143
Uppercase Greek Letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1144
Special Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1145

Overview
The ENTRY, ENTRYTITLE, and ENTRYFOOTNOTE statements all use the same
syntax for specifying one or more pieces of text called text items. For example, here is
the ENTRYTITLE syntax:
ENTRYTITLE text-item <…<text-item>> </option(s)>;
Each text item can be specified using the following syntax:
<prefix-option…<prefix-option>> "string" | dynamic | character-expression | {text-command}
The simplest specification for each statement is to specify a string in quotation marks.
For example, you might specify a graph title as follows:

entrytitle "Height and Weight by Sex";

In this example, the string “Height and Weight by Sex” is formatted as the graph title
and displayed in the title position. If multiple strings are specified, then they are
concatenated into a single line of text. This second specification generates the same title:

entrytitle "Height " "and " "Weight " "by " "Sex";
1138 Chapter 10 • Managing Text Items

To provide control over the text, multiple prefix options can precede each text item, and
the text item can be a string literal, a dynamic, or a text command. All text items with the
same HALIGN= setting are concatenated into one string. Up to three strings with
different horizontal alignment can result for each ENTRY statement. Leading and
trailing blanks in the concatenated string are always used.
• a string must be enclosed in quotation marks.
• a character expression must be enclosed in an EVAL function.
• a text command must be enclosed in braces. (See “Using Text Commands” on page
1140).

Using Prefix Options

Available Prefix Options

The following prefix options are available on ENTRY, ENTRYTITLE, and
ENTRYFOOTNOTE statements:
HALIGN
specifies the horizontal alignment of a text item.
TEXTATTRS
specifies font attributes for a text item.
When used, a prefix option applies not only to immediately following piece of text but
also to ALL subsequent text strings and text-commands. If the same prefix option
appears more than once, then each use overrides the last used prefix of the same name.

Managing Horizontal Alignment

For the ENTRY, ENTRYTITLE, and ENTRYFOOTNOTE statements, the default
horizontal alignment is CENTER.

entry "One" " Two" " Three" " Four" " Five" " Six";

The HALIGN= option can be used to change the horizontal alignment to LEFT,
CENTER, or RIGHT. The following specification left-justifies the text:

entry halign=left
"One" " Two" " Three" " Four" " Five" " Six";

Text items are positionally concatenated by alignment area. For example, the following
specification left-justifies the first three strings and right-justifies the last three strings:
Using Prefix Options 1139

entry halign=left "One" " Two" " Three"

halign=right "Four" " Five" " Six";

Even if the HALIGN= specifications are jumbled, the final text is nevertheless
positionally concatenated by alignment area:

entry halign=right "Five" halign=left "One"

halign=right " Six" halign=center "Three"
halign=left " Two" halign=center " Four";

Note: When long strings are used or the bounding container is constrained, the
alignment areas might overlap.

Managing Font Attributes

For the ENTRY, ENTRYTITLE, and ENTRYFOOTNOTE statements, the default font
attributes are determined by the active ODS style. The TEXTATTRS= option can be
used to change the default font attributes. The TEXTATTRS= option syntax is as follows
(see “General Syntax for Attribute Options” on page 1347 for the syntax on using a
style element and “Text Options” on page 1351 for available text options):
TEXTATTRS=style-element | style-element(text-options) | (text-options)
For example, the following ENTRYTITLE statement uses prefix options to create this
title line:

entrytitle textattrs=(color=black) "Center"

textattrs=(color=red) " Text"
halign=right textattrs=(color=black size=10pt) "Right "
textattrs=(color=red size=10pt) "side"
halign=left
textattrs=(color=black style=italic size=10pt) "Left"
textattrs=(color=red style=italic size=10pt) " side";

• The TEXTATTRS= options are reset each time a new TEXTATTRS= appears—there
is no “carry over” or accumulation of the settings.
• Blanks (spaces) must be provided as needed to achieve the desired concatenation.
When concatenating dynamics that are stripped of leading and trailing blanks, a
literal space must be inserted to separate them, as shown in the following example:

entry _DYN1 " " _DYN2;

Alternatively, font attributes can be specified by overriding the style element defaults.
The following specification overrides the settings of the GraphTitleText style element,
which sets the default attributes for ENTRYTITLE text:

entrytitle
halign=left
textattrs=GraphTitleText(color=black style=italic) "Left"
textattrs=GraphTitleText(color=red style=italic) " side"
1140 Chapter 10 • Managing Text Items

halign=center
textattrs=(color=black) "Center"
textattrs=(color=red) " Text"
halign=right
textattrs=GraphTitleText(color=black weight=bold) "Right "
textattrs=GraphTitleText(color=red weight=bold) "side" ;

Dynamics can also be used in the text strings. In the following ENTRYTITLE statement,
assume that _DEPLABEL and _MODELLABEL are dynamics that are specified on
PROC TEMPLATE’s DYNAMIC statement:

entrytitle "Residual by Predicted for " _DEPLABEL

halign=left textattrs=GraphTitleText _MODELLABEL /
pad=(bottom=5);

• The default style element for ENTRYTITLE is GraphTitleText, so all three text items
(one literal and two dynamics) get these font properties as a starting point.
• The text "Residual by Predicted for " _DEPLABEL is center-aligned by default.
• The prefix options HALIGN= and TEXTATTRS= override the center alignment and
font properties for the text _MODELLABEL.
This could have been coded as follows:

entrytitle halign=left
textattrs=GraphTitleText _MODELLABEL
halign=center textattrs=()
"Residual by Predicted for " _DEPLABEL;

In this example, the second HALIGN= and TEXTATTRS=() are necessary to reset
alignment and font properties to their defaults.
The string length of the resolved dynamic _MODELLABEL does not affect the
placement of the center-aligned text.

Using Text Commands

Available Text Commands

Text commands on ENTRY, ENTRYTITLE, and ENTRYFOOTNOTE statements are
special in-line instructions that either modify the appearance of the text or script special
characters. The following text commands are available:
{SUB}
specifies that the string or dynamic appears as a subscript.
{SUP}
specifies that the string or dynamic appears as a superscript.
{UNICODE}
specifies a glyph (graphical character) to be displayed using its Unicode specification
or keyword equivalent.
The general form of a text command is
{command argument(s)}
Using Text Commands 1141

The opening and closing braces are required to denote the scope of the command.

Subscripting and Superscripting Text

The {SUB} and {SUP} text commands are used to subscript and superscript text. Each
of these text commands accepts a string or a dynamic for its argument(s).
In the following example, if _RSQUARE is an existing dynamic that resolves to the
value 0.7434, then the following ENTRY specification superscripts the string value "2"
to generate this text:

entry textattrs=(weight=bold) "R" {sup "2"}

textattrs=() "=" _RSQUARE;

The textattrs=() option cancels all style overrides and reverts to the default text
properties.

Using UNICODE Text

The {UNICODE} text command places special characters into the text and accepts any
of the following for its argument(s):
• a hexadecimal Unicode Code Point for a character (for example, "03B1"x)
• a reserved keyword for a commonly used code point (for example, BETA)
• a dynamic that resolves to a hexadecimal value or keyword.
The tables in “Reserved Keywords and Unicode Values” on page 1142 provide a list of
the commonly used reserved keywords and Unicode values (the tables are not complete,
but they provide an idea about what is possible).
Multiple arguments can be used within the scope of a single UNICODE text command.
For example, the following two specifications are equivalent:

{unicode "03b1"x beta}

{unicode "03b1"x} {unicode beta}

The default formatting for the UNICODE text is derived from the GraphUnicodeText
style element.
In the following example, if _ALPHAVAL is an existing dynamic that resolves to the
value 0.05, then the following ENTRY specification generates this text:

entry {unicode alpha} " = " _ALPHAVAL;

By combining the TEXTATTRS= prefix option with the {SUB} and {UNICODE} text
commands, you can generate the following text:
1142 Chapter 10 • Managing Text Items

Entry textattrs=(style=italic) "E(Y)" textattrs=() " = "

{unicode beta} {sub "0"} " + "
{unicode beta} {sub "1"} "x" {sub "1"}
" + " {unicode beta} {sub "2"} "x" {sub "2"};

Rules for Unicode and Special Character Specifications

The following rules apply to Unicode and special character specifications in ODS
graphics:
• Each character can be specified by looking up its code and specifying it as a
hexadecimal constant. Example: {unicode ’221e’x}.
• Lowercase Greek letters can be specified by using names instead of hexadecimal
constants. Example: {unicode alpha}.
• Uppercase Greek letters can be specified by using names followed by _u instead of a
hexadecimal constants. Example: {unicode alpha_u}.
• Superscript and subscript have special abbreviations. Examples: {sup 2} and {sub
2}.
• The {SUP} and {SUB} specifications must not appear escaped and in quotation
marks in the GTL. They must appear outside of quotation marks.
• Some characters overprint the character that comes before. Example: ’El nin’
{tilde} ’o’, which is equivalent to ’El nin’ {unicode ’0303’x} ’o’
creates ‘El niño’.
• Specifications inside quotation marks are escaped. Example: "(*ESC*){unicode
beta}".
• Specifications outside quotation marks are not escaped. Example: {unicode
beta}.
For more information about using text throughout the GTL (for example, using Unicode
values in labels), see SAS Graph Template Language: User's Guide.

Reserved Keywords and Unicode Values

Overview
The tables in this section show some of the reserved keywords and Unicode values that
can be used with the UNICODE text command. For information about rendering
Unicode characters, see “Managing the String on Text Statements” in SAS Graph
Template Language: User's Guide.
Note the following:
• Keywords and Unicode values are not case-sensitive: "03B1"x is the same code point
as "03b1"x.
• The word blank is the keyword for a blank space.
Reserved Keywords and Unicode Values 1143

Lowercase Greek Letters

Keyword Glyph Unicode Description

alpha α 03B1 lowercase alpha

beta β 03B2 lowercase beta

gamma γ 03B3 lowercase gamma

delta δ 03B4 lowercase delta

epsilon ε 03B5 lowercase epsilon

zeta ζ 03B6 lowercase zeta

eta η 03B7 lowercase eta

theta θ 03B8 lowercase theta

iota ι 03B9 lowercase iota

kappa κ 03BA lowercase kappa

lambda λ 03BB lowercase lambda

mu μ 03BC lowercase mu

nu ν 03BD lowercase nu

xi ξ 03BE lowercase xi

omicron ο 03BF lowercase omicron

pi π 03C0 lowercase pi

rho ρ 03C1 lowercase rho

sigma σ 03C3 lowercase sigma

tau τ 03C4 lowercase tau

upsilon υ 03C5 lowercase upsilon

phi φ 03C6 lowercase phi

chi χ 03C7 lowercase chi

psi ψ 03C8 lowercase psi

1144 Chapter 10 • Managing Text Items

Keyword Glyph Unicode Description

omega ω 03C9 lowercase omega

Uppercase Greek Letters

Table 10.1 Uppercase Greek Letters

Keyword Glyph Unicode Description

alpha_u Α 0391 uppercase alpha

beta_u Β 0392 uppercase beta

gamma_u Γ 0393 uppercase gamma

delta_u Δ 0394 uppercase delta

epsilon_u Ε 0395 uppercase epsilon

zeta_u Ζ 0396 uppercase zeta

eta_u Η 0397 uppercase eta

theta_u Θ 0398 uppercase theta

iota_u Ι 0399 uppercase iota

kappa_u Κ 039A uppercase kappa

lambda_u Λ 039B uppercase lambda

mu_u Μ 039C uppercase mu

nu_u Ν 039D uppercase nu

xi_u Ξ 039E uppercase xi

omicron_u Ο 039F uppercase omicron

pi_u Π 03A0 uppercase pi

rho_u Ρ 03A1 uppercase rho

sigma_u Σ 03A3 uppercase sigma

tau_u Τ 03A4 uppercase theta

upsilon_u Υ 03A5 uppercase upsilon

Reserved Keywords and Unicode Values 1145

Keyword Glyph Unicode Description

phi_u Φ 03A6 uppercase phi

chi_u Χ 03A7 uppercase chi

psi_u Ψ 03A8 uppercase psi

omega_u Ω 03A9 uppercase omega

Special Characters

Keyword Glyph Unicode Description

prime ´ 00B4 single prime sign

bar ̅ 0305 combining overline*

bar2 ̿ 033F combining double

overline*

tilde ̃ 0303 combining tilde*

hat ̂ 0302 combining

circumflex accent*

* This is an overstriking character that requires a Unicode font to render properly.

1146 Chapter 10 • Managing Text Items
1147

Chapter 11
Text Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1147
ENTRY Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1147
ENTRYFOOTNOTE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1154
ENTRYTITLE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1162

Dictionary

ENTRY Statement
Displays a line of text in the graphical area.
Requirement: An ENTRY statement must be specified within a LAYOUT, HEADER, SIDEBAR, or
CELL statement block.

Syntax
ENTRY text-item <text-item …> </option(s)>;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the text.
OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the entry border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry border.
1148 Chapter 11 • Text Statements

ROTATE=0 | 90 | 180 | 270

specifies the angle of text rotation measured in degrees.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
as a statement option, specifies the properties of the text. As a prefix-option,
specifies the properties of individual text-items.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the text is automatically aligned within its parent when
nested within an overlay-type layout.
VALIGN=CENTER | TOP | BOTTOM
specifies the vertical alignment of the text.

Required Argument
text-item <…<text-item>>
specifies one or more pieces of text for the entry. Each text-item has the following
form:
<prefix-option <… prefix-option>>"string" | dynamic | character-expression |
{text-command}
Each piece of text can have multiple prefix settings that precede the piece of text. A
piece of text is either a string literal, a dynamic, or a text command. All text-items
with the same HALIGN= setting are concatenated into one string. Up to three strings
can result for each ENTRY statement. Leading and trailing blanks in the
concatenated string are always used.
When a prefix option is used, it applies not only to the text that immediately follows
the prefix option, but also to all subsequent text strings and text commands. If the
same prefix option appears more than once, then that prefix overrides the last used
prefix of the same name for the subsequent text strings and text commands.

Restriction The maximum length for the entry text is 512 characters in SAS 9.4
and earlier releases. Starting with the first maintenance release of
SAS 9.4, this restriction is removed.

Requirements string must be enclosed in quotation marks.

character-expression must be enclosed in an EVAL function.

text-command must be enclosed in braces.

Note Leading spaces are preserved, and trailing spaces are stripped.

See Chapter 10, “Managing Text Items,” on page 1137

Optional Arguments
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the text is automatically aligned within its parent when nested
within an overlay-type layout. For more information about how child positions are
determined in an overlay-type layout, see the “LAYOUT OVERLAY Statement” on
page 142.
ENTRY Statement 1149

NONE
does not automatically align the text within the area. Alignment is set with
HALIGN= and VALIGN= options.
AUTO
within the parent layout, attempts to center the text in the area that is farthest
from any surrounding data point markers.

Restriction This option is available only if the parent layout contains a scatter
plot. Otherwise, it is ignored.

(location-list)
within the available area, restricts the text’s possible locations to those locations
in the specified location-list, and uses the location-list position that least collides
with the other graphics features in the area. The location-list is space-separated
and can contain any of these locations: TOPLEFT, TOP, TOPRIGHT, LEFT,
CENTER, RIGHT, BOTTOMLEFT, BOTTOM, and BOTTOMRIGHT.
Example: (TOPRIGHT TOPLEFT)

Default NONE

Requirement For this option to take effect, the ENTRY statement must be in a
LAYOUT OVERLAY or LAYOUTOVERLAYEQUATED block.

Interaction When AUTOALIGN= is not NONE and the parent layout is an

overlay-type layout, the HALIGN= and the VALIGN= options are
ignored.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect for the color to be seen. By default,
OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the text.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

1150 Chapter 11 • Text Statements

OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the entry border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the entry border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the entry border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
ENTRY Statement 1151

TOP=dimension specifies the amount of extra space added to the

top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default (LEFT=3 RIGHT=3 TOP=0 BOTTOM=0)

Note The default units for dimension are pixels.

See “dimension” on page 1340

ROTATE=0 | 90 | 180 | 270

specifies the angle of text rotation measured in degrees. The angle is measured from
a horizontal line passing through the middle of the bounding box of the text, counter-
clockwise starting at the center of the bounding box.

Default 0. No rotation is performed.

Restriction Only angles of 0, 90, 180, or 270 degrees are allowed.

Interaction The bounding box is the determined by the size of the text in the
current font plus any horizontal and vertical padding. See
TEXTATTRS= and PAD= .

TEXTATTRS=style-element | style-element (text-options) | (text-options)

as a statement option, specifies the properties of the text. As a prefix-option, specifies
the properties of individual text-items.

Default The GraphValueText style element.

Notes This option can be used as both a prefix option and a statement option.
When used as a prefix option, it overrides the statement option.

When used as a prefix option, TEXTATTRS= cancels the last used

TEXTATTRS= prefix option. It resets all text options to those set by 1) the
TEXTATTRS= statement option or 2) the default style element for the
statement (GraphValueText) if no TEXTATTRS= statement option is used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

VALIGN=CENTER | TOP | BOTTOM

specifies the vertical alignment of the text.

Default CENTER

Interaction This option is ignored when AUTOALIGN= is not NONE and the
parent layout is an overlay-type layout.
1152 Chapter 11 • Text Statements

Prefix Options
HALIGN=CENTER | LEFT | RIGHT
prefix-option that specifies the horizontal alignment of a text-item. Each text-item has
a horizontal alignment, and text-items with the same alignment are always grouped
together.

Default CENTER

Interaction This option is ignored when AUTOALIGN= is not NONE and the
parent layout is an overlay-type layout.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

See “TEXTATTRS=style-element | style-element (text-options) | (text-options) ” on
page 1151

Text Commands
{ SUB "string" | dynamic }
text-command that specifies that the string or dynamic is to appear as subscript text.

See “Rules for Unicode and Special Character Specifications” on page 1142

Example entry "y = " b{sub "0"} " + b" {sub "1"} "x";

{ SUP "string" | dynamic }

text-command that specifies that the string or dynamic is to appear as superscript
text.

See “Rules for Unicode and Special Character Specifications” on page 1142

Example entry "R" {sup "2"} " = " {format (6.4) RSQUARED} ;

{ UNICODE "hex-string"x | keyword | dynamic }

text-command that specifies a glyph (character) to be displayed using its Unicode
specification or keyword equivalent.
"hex-string"x
a four-byte hexadecimal constant that represents a UNICODE character in the
current font. For a complete listing, see http://unicode.org/charts/charindex.html.
keyword
a SAS keyword for a UNICODE character. For a listing of SAS supplied
keywords, see “Reserved Keywords and Unicode Values” on page 1142.
dynamic
a dynamic variable that resolves to either "hex-string"x or a keyword for a
UNICODE character.

Note The UNICODE text command attempts to access a UNICODE value in

the current font. Not all fonts support accessing characters through their
UNICODE value. Some fonts support only a limited set of UNICODE
values. If the UNICODE value is not accessible, then the command might
be ignored or a nonprintable character might be substituted.

See “Using UNICODE Text” on page 1141

“Rules for Unicode and Special Character Specifications” on page 1142

Example: ENTRY Statement 1153

Example entry {unicode alpha} "=" CONF;

entry {unicode "03B1"x} "=" CONF;

Details
An ENTRY statement creates one line of text in the plot area. The statement must be
specified within a LAYOUT, HEADER, SIDEBAR, or CELL statement block. It cannot
be specified outside of one of these blocks, where global statements like ENTRYTITLE
and ENTRYFOOTNOTE are used.

The text line specified in an ENTRY statement can be made of several pieces of the text
called text-items. Statement options that are used establish properties for the entire line
of text (that is, all text-items). These properties can be overridden with prefix-options
that are specific to one or more text-items. See “Required Argument” on page 1148 for
more information.
You can specify an ENTRY statement inside or outside of a nested statement block:
• When you specify an ENTRY statement inside a nested LAYOUT, HEADER,
SIDEBAR, or CELL statement block, then, by default, the text is placed inside the
bounding area of the results that the nested statement block creates.
• When you specify an ENTRY statement outside of a nested LAYOUT, HEADER,
SIDEBAR, or CELL statement block, then the text is placed outside of the bounding
area of the results that the nested statement block creates.

Example: ENTRY Statement

The following graph was generated by the “Example Program” on page 1154:
1154 Chapter 11 • Text Statements

Example Program
proc template;
define statgraph entry;
begingraph;
layout overlay;

entry halign=right "First entry statement" /

valign=top;

histogram weight;

entry halign=right "Second entry statement";

entry halign=right "Third entry statement" /

valign=bottom pad=(bottom=40px);

endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=entry;

run;

ENTRYFOOTNOTE Statement
Displays a footnote.
ENTRYFOOTNOTE Statement 1155

Syntax
ENTRYFOOTNOTE text-item <text-item …> </option(s)>;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the text.
OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the entry border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry border.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
as a statement option, specifies the properties of the text. As a prefix-option,
specifies the properties of individual text-items.

Text options
HALIGNCENTER=AUTO | GRAPH
specifies whether the text is centered automatically by the system or is
always centered in the graph area.
SHORTTEXT=(text-item <...text-item>)
specifies alternate text to use if the specified text is too long for the output
width.
TEXTFITPOLICY=WRAP | SHORT | TRUNCATE
specifies how to handle text that is too long to fit in the output width.

Required Argument
text-item <…text-item>
specifies one or more pieces of text for the footnote. Each text-item has the following
form:
<prefix-option …<prefix-option>>"string" | dynamic | character-expression |
{text-command}
Each piece of text can have multiple prefix options that precede the piece of text. A
piece of text is either a string literal, a dynamic, or a text command. All text-items
with the same HALIGN= setting are concatenated into one string. Up to three strings
can result for each ENTRY statement. Leading and trailing blanks in the
concatenated string are always used.
When a prefix option is used, it applies not only to the text that immediately follows
the prefix option, but also to all subsequent text strings and text commands. If the
same prefix option appears more than once, then that prefix overrides the last used
prefix of the same name for the subsequent text strings and text commands.
1156 Chapter 11 • Text Statements

Restriction The maximum length for the footnote text is 512 characters in SAS
9.4 and earlier releases. Starting with the first maintenance release
of SAS 9.4, this restriction is removed.

Requirements string must be enclosed in quotation marks.

character-expression must be enclosed in an EVAL function.

text-command must be enclosed in braces.

Note Leading spaces are preserved, and trailing spaces are stripped.

See Chapter 10, “Managing Text Items,” on page 1137

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect for the color to be seen. By default,
OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the text.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

HALIGNCENTER=AUTO | GRAPH
specifies whether the text is centered automatically by the system or is always
centered in the graph area.
AUTO
specifies that the system determines how the text is centered, as follows:
• For LAYOUT GRIDDED, LAYOUT OVERLAY3D, and LAYOUT
REGION layouts, and for LAYOUT LATTICE, LAYOUT DATAPANEL,
and LAYOUT DATALATTICE layouts that have more than one column,
center the text in the graph area.
ENTRYFOOTNOTE Statement 1157

• For LAYOUT OVERLAY and LAYOUT OVERLAYEQUATED layouts, and

for LAYOUT LATTICE, LAYOUT DATAPANEL, and LAYOUT
DATALATTICE layouts that have only one column, center the text in the
graph wall area. If the length of the text exceeds the width of the graph wall
area, then center the text in the graph area instead.
GRAPH
specifies that the text is always centered in the graph area.

Default AUTO

Interaction The prefix option HALIGN= must specify CENTER for this option to
have any effect.

OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the entry border.
AUTO
specifies that the default outside padding for this component is used.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the entry border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

1158 Chapter 11 • Text Statements

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the entry border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default (LEFT=3 RIGHT=3 TOP=0 BOTTOM=0)

Note The default units for dimension are pixels.

See “dimension” on page 1340

SHORTTEXT=(text-item <...text-item>)
specifies alternate text to use if the specified text is too long for the output width. If
the shortened text is itself too long, then it is truncated.

Interactions This option has no effect unless TEXTFITPOLICY= SHORT.

This option is ignored if any text-items include an HALIGN= prefix

option.

The horizontal alignment of the shortened text is derived from the

horizontal alignment of the text to be shortened.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

as a statement option, specifies the properties of the text. As a prefix-option, specifies
the properties of individual text-items.

Default The GraphFootnoteText style element.

Notes This option can be used as both a prefix option and a statement option.
When used as a prefix option, it overrides the statement option.

When used as a prefix option, TEXTATTRS= cancels the last used

TEXTATTRS= prefix option. It resets all text options to those set by 1) the
TEXTATTRS= statement option or 2) the default style element for the
statement (GraphValueText) if no TEXTATTRS= statement option is used.
ENTRYFOOTNOTE Statement 1159

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TEXTFITPOLICY=WRAP | SHORT | TRUNCATE

specifies how to handle text that is too long to fit in the output width.
WRAP
specifies that the text wraps to the next line(s).

Restriction Text wrapping is available only for ENTRYFOOTNOTE

statements that appear in the outermost layout.

SHORT
specifies that the text indicated by the SHORTTEXT= option be substituted when
the text does not fit.

Requirement You must specify the SHORTTEXT= option for this option to
take effect.

Note If the short text is also too long, then it is truncated.

TRUNCATE
specifies that the text is truncated to make it fit.

Default WRAP

Prefix Options
HALIGN=CENTER | LEFT | RIGHT
prefix-option that specifies the horizontal alignment of a text-item. Each text-item has
a horizontal alignment, and text-items with the same alignment are always grouped
together.

Default CENTER

TEXTATTRS=style-element | style-element (text-options) | (text-options)

See “TEXTATTRS=style-element | style-element (text-options) | (text-options) ” on
page 1158

Text Commands
{ SUB "string" | dynamic }
text-command that specifies that the string or dynamic is to appear as subscript text.

See “Rules for Unicode and Special Character Specifications” on page 1142

Example entryfootnote "y = " b{sub "0"} " + b" {sub "1"} "x";

{ SUP "string" | dynamic }

text-command that specifies that the string or dynamic is to appear as superscript
text.

See “Rules for Unicode and Special Character Specifications” on page 1142
1160 Chapter 11 • Text Statements

Example entryfootnote "R" {sup "2"} " = " RSQUARED;

{ UNICODE "hex-string"x | keyword | dynamic }

text-command that specifies a glyph (character) to be displayed using its Unicode
specification or keyword equivalent.
"hex-string"x
a four-byte hexadecimal constant that represents a UNICODE character in the
current font. For a complete listing, see http://unicode.org/charts/charindex.html.
keyword
a SAS keyword for a UNICODE character. For a listing of SAS supplied
keywords, see “Reserved Keywords and Unicode Values” on page 1142.
dynamic
a dynamic variable that resolves to either "hex-string"x or a keyword for a
UNICODE character.

Note The UNICODE text command attempts to access a UNICODE value in

the current font. Not all fonts support accessing characters through their
UNICODE value. Some fonts support only a limited set of UNICODE
values. If the UNICODE value is not accessible, then the command might
be ignored or a nonprintable character might be substituted.

See “Using UNICODE Text” on page 1141

“Rules for Unicode and Special Character Specifications” on page 1142

Example entryfootnote {unicode alpha} "=" CONF;

entryfootnote {unicode "03B1"x} "=" CONF;

Details
The ENTRYFOOTNOTE statement places footnote text below the graphical area. More
than one ENTRYFOOTNOTE statement can be used. Footnotes appear in the order of
the ENTRYFOOTNOTE statements within the template.

When adding an ENTRYFOOTNOTE statement to a template definition, the statement

must be located within the BEGINGRAPH block but outside of the outermost layout
block.
• All ENTRYFOOTNOTE statements that appear in the template are gathered and
placed in the ENTRYFOOTNOTE area.
• The placement of an ENTRYFOOTNOTE statement is relevant only in relation to
other ENTRYFOOTNOTE statements.
• As the number of ENTRYFOOTNOTE statements increases the size of the graphical
area is reduced.
Example: ENTRYFOOTNOTE Statement 1161

Footnotes always span the entire width of the output. By default, footnotes are “center-
aligned,” based on the type of the outermost layout. The meaning of “center-aligned”
varies by layout type and the number of columns in the layout:

Layout Type Default horizontal centering of footnotes

GRIDDED Centered on width of entire graph

OVERLAY3D
LATTICE (COLUMNS > 1)
DATAPANEL (COLUMNS > 1)
DATALATTICE (COLUMNS > 1)

OVERLAY Centered on the plot area

OVERLAYEQUATED
LATTICE (COLUMNS=1)
DATAPANEL (COLUMNS= 1)
DATALATTICE (COLUMNS = 1)

Example: ENTRYFOOTNOTE Statement

The following graph was generated by the “Example Program” on page 1162:
1162 Chapter 11 • Text Statements

Example Program
proc template;
define statgraph entryfootnote;
begingraph;

entryfootnote "First entryfootnote statement" ;

layout overlay;
histogram weight;
endlayout;

entryfootnote "Second entryfootnote statement" ;

entryfootnote "Third entryfootnote statement" ;

endgraph;
end;
run;

proc sgrender data=sashelp.cars template=entryfootnote;

run;

ENTRYTITLE Statement
Displays a title.

Syntax
ENTRYTITLE text-item <text-item …> </option(s)>;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the text.
OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the entry border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry border.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
as a statement option, specifies the properties of the text. As a prefix-option,
specifies the properties of individual text-items.

Text options
HALIGNCENTER=AUTO | GRAPH
ENTRYTITLE Statement 1163

specifies whether the text is centered automatically by the system or is

always centered in the graph area.
SHORTTEXT=(text-item <…text-item>)
specifies alternate text to use if the specified text is too long for the output
width. If the shortened text is itself too long, then it is truncated.
TEXTFITPOLICY=WRAP | SHORT | TRUNCATE
specifies how to handle text that is too long to fit in the output width.

Required Argument
text-item <text-item …>
specifies one or more pieces of text for the entry. Each text-item has the following
form:
<prefix-option <…prefix-option>>"string" | dynamic | character-expression |
{text-command}
Each piece of text can have multiple prefix options that precede the piece of text. A
piece of text is either a string literal, a dynamic, or a text command. All text-items
with the same HALIGN= setting are concatenated into one string. Up to three strings
can result for each ENTRY statement. Leading and trailing blanks in the
concatenated string are always used.
When a prefix option is used, it applies not only to the text that immediately follows
the prefix option, but also to all subsequent text strings and text commands. If the
same prefix option appears more than once, then that prefix overrides the last used
prefix of the same name for the subsequent text strings and text commands.

Restriction The maximum length for the title text is 512 characters in SAS 9.4
and earlier releases. Starting with the first maintenance release of
SAS 9.4, this restriction is removed.

Requirements string must be enclosed in quotation marks.

character-expression must be enclosed in an EVAL function.

text-command must be enclosed in braces.

Note Leading spaces are preserved, and trailing spaces are stripped.

See Chapter 10, “Managing Text Items,” on page 1137 for more
information and several examples.

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect for the color to be seen. By default,
OPAQUE=FALSE.
1164 Chapter 11 • Text Statements

BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.

Default FALSE

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the text.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

HALIGNCENTER=AUTO | GRAPH
specifies whether the text is centered automatically by the system or is always
centered in the graph area.
AUTO
specifies that the system determines how the text is centered, as follows:
• For LAYOUT GRIDDED, LAYOUT OVERLAY3D, and LAYOUT
REGION layouts, and for LAYOUT LATTICE, LAYOUT DATAPANEL,
and LAYOUT DATALATTICE layouts that have more than one column,
center the text in the graph area.
• For LAYOUT OVERLAY and LAYOUT OVERLAYEQUATED layouts, and
for LAYOUT LATTICE, LAYOUT DATAPANEL, and LAYOUT
DATALATTICE layouts that have only one column, center the text in the
graph wall area. If the length of the text exceeds the width of the graph wall
area, then center the text in the graph area instead.
GRAPH
specifies that the text is always centered in the graph area.

Default AUTO

Interaction The prefix option HALIGN= must specify CENTER for this option to
have any effect.

OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent (FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page 1339 for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the entry border.
AUTO
specifies that the default outside padding for this component is used.
ENTRYTITLE Statement 1165

dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the entry border.
(pad-options)
a space-separated list of one or more of the following name-value pair options,
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space to add to the

left side.
RIGHT=dimension specifies the amount of extra space to add to the
right side.
TOP=dimension specifies the amount of extra space to add to the
top.
BOTTOM=dimension specifies the amount of extra space to add to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page 1340

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the entry border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default (LEFT=3 RIGHT=3 TOP=0 BOTTOM=0)

Note The default units for dimension are pixels.

1166 Chapter 11 • Text Statements

See “dimension” on page 1340

SHORTTEXT=(text-item <…text-item>)
specifies alternate text to use if the specified text is too long for the output width. If
the shortened text is itself too long, then it is truncated.

Interactions This option has no effect unless TEXTFITPOLICY= SHORT.

This option is ignored if any text-items include an HALIGN= prefix

option.

The horizontal alignment of the shortened text is derived from the

horizontal alignment of the title to be shortened.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

as a statement option, specifies the properties of the text. As a prefix-option, specifies
the properties of individual text-items.

Default The GraphTitleText style element.

Notes This option can be used as both a prefix option and a statement option.
When used as a prefix option, it overrides the statement option.

When used as a prefix option, TEXTATTRS= cancels the last used

TEXTATTRS= prefix option. It resets all text options to those set by 1) the
TEXTATTRS= statement option or 2) the default style element for the
statement (GraphValueText) if no TEXTATTRS= statement option is used.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

TEXTFITPOLICY=WRAP | SHORT | TRUNCATE

specifies how to handle text that is too long to fit in the output width.
WRAP
specifies that the text wraps to the next line(s).

Restriction Text wrapping is available only for ENTRYTITLE statements that

appear in the outermost layout.

SHORT
specifies that the text indicated by the SHORTTEXT= option be substituted when
the title does not fit.

Requirement You must specify the SHORTTEXT= option for this option to
take effect.

Note If the short text is also too long, then it is truncated.

TRUNCATE
specifies that the text is truncated to make it fit.

Default WRAP
ENTRYTITLE Statement 1167

Prefix Options
HALIGN=CENTER | LEFT | RIGHT
prefix-option that specifies the horizontal alignment of a text-item. Each text-item has
a horizontal alignment, and text-items with the same alignment are always grouped
together.

Default CENTER

TEXTATTRS=style-element | style-element (text-options) | (text-options)

See “TEXTATTRS=style-element | style-element (text-options) | (text-options) ” on
page 1166

Text Commands
{ SUB "string" | dynamic }
text-command that specifies that the string or dynamic is to appear as subscript text.

See “Rules for Unicode and Special Character Specifications” on page 1142

Example entrytitle "y = " b{sub "0"} " + b" {sub "1"} "x";

{ SUP "string" | dynamic }

text-command that specifies that the string or dynamic is to appear as superscript
text.

See “Rules for Unicode and Special Character Specifications” on page 1142

Example entrytitle "R" {sup "2"} " = " RSQUARED;

{ UNICODE "hex-string"x | keyword | dynamic }

text-command that specifies a glyph (character) to be displayed using its Unicode
specification or keyword equivalent.
"hex-string"x
a four-byte hexadecimal constant that represents a UNICODE character in the
current font. For a complete listing, see http://unicode.org/charts/charindex.html.
keyword
a SAS keyword for a UNICODE character. For a listing of SAS supplied
keywords, see “Reserved Keywords and Unicode Values” on page 1142.
dynamic
a dynamic variable that resolves to either "hex-string"x or a keyword for a
UNICODE character.

Note The UNICODE text command attempts to access a UNICODE value in

the current font. Not all fonts support accessing characters through their
UNICODE value. Some fonts support only a limited set of UNICODE
values. If the UNICODE value is not accessible, then the command might
be ignored or a nonprintable character might be substituted.

See “Using UNICODE Text” on page 1141

“Rules for Unicode and Special Character Specifications” on page 1142

Example entrytitle {unicode alpha} "=" CONF;

entrytitle {unicode "03B1"x} "=" CONF;
1168 Chapter 11 • Text Statements

Details
The ENTRYTITLE statement places title text above the graphical area. More than one
ENTRYTITLE statement can be used. Titles appear in the order of the ENTRYTITLE
statements within the template.

When adding an ENTRYTITLE statement to a template definition, the statement must be

located within the BEGINGRAPH block but outside of the outermost layout block.
• All ENTRYTITLE statements that appear in the template are gathered and placed in
the ENTRYTITLE area.
• The placement of an ENTRYTITLE statement is relevant only in relation to other
ENTRYTITLE statements.
• As the number of ENTRYTITLE statements increases the size of the graphical area
is reduced.
Titles always span the entire width of the output. By default, titles are “center-aligned,”
based on the type of the outermost layout. The meaning of “center-aligned” varies by
layout type and the number of columns in the layout:

Layout Type Default horizontal centering of titles

GRIDDED Centered on width of entire graph

OVERLAY3D
LATTICE (COLUMNS > 1)
DATAPANEL (COLUMNS > 1)
DATALATTICE (COLUMNS > 1)

OVERLAY Centered on the plot area

OVERLAYEQUATED
LATTICE (COLUMNS=1)
DATAPANEL (COLUMNS= 1)
DATALATTICE (COLUMNS = 1)

Example: ENTRYTITLE Statement

The following graph was generated by the “Example Program” on page 1169:
Example: ENTRYTITLE Statement 1169

Example Program
proc template;
define statgraph entrytitle;
begingraph;

entrytitle "First entrytitle statement" ;

layout overlay;
histogram weight;
endlayout;

entrytitle "Second entrytitle statement" ;

entrytitle "Third entrytitle statement" ;

endgraph;
end;
run;

proc sgrender data=sashelp.cars template=entrytitle;

run;
1170 Chapter 11 • Text Statements
1171

Part 8

Custom Marker Definition

Statements

Chapter 12
Custom Marker Definition Statements . . . . . . . . . . . . . . . . . . . . . . . . . . 1173
1172
1173

Chapter 12
Custom Marker Definition
Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1173
SYMBOLCHAR Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1173
SYMBOLIMAGE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1180

Dictionary

SYMBOLCHAR Statement
Defines a marker symbol using a Unicode character so that the symbol can be referenced in other
statements.
Note: This statement is valid in the first maintenance release of SAS 9.4 and later
releases.

Syntax
SYMBOLCHAR NAME=marker-name CHAR="hex-string"x | keyword | dynamic
</ option(s)>;

Required Arguments
NAME=marker-name
specifies a name for the marker symbol. The name can be used in statements that
support marker symbols.

Interaction If the name matches one of the system-provided symbol names such as
CIRCLE, then the system symbol is replaced by the user-defined
symbol. See “Marker Options” on page 1350 for a list of the system-
provided symbols.

Note Do not enclose the name in quotation marks.

See “Details” on page 1175

CHAR="hex-string"x | keyword | dynamic

specifies a glyph (character) to be used as the marker symbol. You specify the
character by using its Unicode specification or its keyword equivalent.
1174 Chapter 12 • Custom Marker Definition Statements

"hex-string"x
specifies a four-byte hexadecimal constant that represents a Unicode character in
the current font. You can find a complete listing of the Unicode hexadecimal
constants at the following URL:http://www.unicode.org/charts/charindex.html
keyword
specifies a SAS keyword for a Unicode character. See Appendix 2, “Reserved
Keywords and Unicode Values,” on page 1343.
dynamic
specifies a reference to a dynamic variable that resolves to either a "hex-string"x
constant or a Unicode character keyword.

Tip This statement attempts to access the specified Unicode value in the current
font. Some fonts do not support accessing characters by using their Unicode
value. Other fonts support only a limited set of Unicode values. If the Unicode
value is not accessible, then this statement might be ignored or a nonprintable
character might be substituted.

Optional Arguments
HOFFSET=number
specifies a horizontal offset for the marker symbol.

Default 0. The marker symbol is centered on its data point.

Range -0.5–+0.5, where 0.5 represents one-half of the original marker size.

Interaction Starting with the third maintenance release of SAS 9.4, the specified
offset is also applied to the marker symbol that is displayed in the
legend.

Notes Prior to the third maintenance release of SAS 9.4, a positive offset
moves the marker symbol bounding box to the right. A negative offset
moves it to the left.

Starting with the third maintenance release of SAS 9.4, a positive offset
moves the marker symbol to the right within its bounding box, and a
negative offset moves it to the left. The bounding box remains centered
on the data point. After the offset, size, and rotation are applied to the
marker symbol, any portion of the symbol that falls outside of the
marker bounding box is clipped.

Tip If clipping occurs, use this option, the VOFFSET= and SCALE=
options, and the suboption SIZE= in the MARKERATTRS= option to
remove the clipping.

ROTATE=number
specifies the angle of rotation for the marker symbol in degrees. Positive angles are
measured in the counter-clockwise direction, and negative angles are measured in the
clockwise direction.

Default 0. No rotation is performed.

Note An angle that exceeds 360 degrees in absolute value can be specified.
SYMBOLCHAR Statement 1175

SCALE=number
specifies a scale factor for the marker symbol as a percentage. The scale factor is
applied to the character's height and width.

Default 1.0 (100%)

Note The outer edges of the image might be clipped by the bounding box when a
large scale factor is specified.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text attributes for the character symbol.

Default The GraphUnicodeText style element.

Restriction Only the text attributes FAMILY=, STYLE=, and WEIGHT= are used.
The COLOR= and SIZE= text attributes are derived from the plot
statement's MARKERATTRS= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

VOFFSET=number
specifies a vertical offset for the marker symbol.

Default 0. The marker symbol is centered on its data point.

Range -0.5–+0.5, where 0.5 represents one-half of the original marker size.

Interaction Starting with the third maintenance release of SAS 9.4, the specified
offset is also applied to the marker symbol that is displayed in the
legend.

Notes Prior to the third maintenance release of SAS 9.4, a positive offset
moves the marker symbol bounding box up. A negative offset moves it
down.

Starting with the third maintenance release of SAS 9.4, a positive offset
moves the marker symbol up within its bounding box, and a negative
offset moves it down. The bounding box remains centered on the data
point. After the offset, size, and rotation are applied to the marker
symbol, any portion of the symbol that falls outside of the marker
bounding box is clipped.

Tip If clipping occurs, use this option, the HOFFSET= and SCALE=
options, and the suboption SIZE= in the MARKERATTRS= option to
remove the clipping.

Details
The SYMBOLCHAR statement defines a custom marker symbol from a Unicode
character. You can use the marker symbol that is created in the following options:
• the DATASYMBOLS= option in the BEGINGRAPH statement
• the SYMBOL= suboption of the MARKERATTRS= option, which is supported by
the following statements:
1176 Chapter 12 • Custom Marker Definition Statements

DISCRETELEGEND (VALUE statement) SCATTERPLOT

LEGENDITEM SCATTERPLOTMATRIX
LINECHART SERIESPLOT
NEEDLEPLOT STEPPLOT
Symbol specifications are not checked for uniqueness. More than one SYMBOLCHAR
statement can define the same character. Therefore, you can use SYMBOLCHAR
statements in IF/ELSE statements. Symbol specifications also are not validated at
compile time. An invalid specification might not generate a warning when the output is
rendered and might create unexpected results.
You can use the COLOR=, SIZE=, and TRANSPARENCY= suboptions of the
MARKERATTRS= option to modify the appearance of markers that are created by the
SYMBOCHAR statement. The WEIGHT= suboption has no effect on these markers.
The markers are clipped to the original marker size after rotation, scaling, and offsets are
applied. If clipping occurs, then you can use the SIZE= suboptions of the
MARKERATTRS= option in the plot statement and the SCALE= option in the
SYMBOLCHAR statement to adjust the size and scaling to eliminate the clipping.

Example: SYMBOLCHAR Statement

This example shows how to create marker symbols from the Unicode ballot X, heavy
character ('2718'x) and check mark, heavy character ('2718'x). It also shows you how to
use the symbols in a scatter plot and how to add them to a legend. In this example, a
grouped bar chart shows the daily failure count for two processes, Process A and Process
B. A scatter plot is overlaid on the bar chart in order to show the total failure count
above each bar. Acceptable counts are indicated by a green Unicode check mark. High
counts are indicated by a red Unicode ballot X. Finally, a legend is used to describe the
custom symbols. The following figure shows the output.
Example: SYMBOLCHAR Statement 1177

Example Program
/* Summarize the data in SASHELP.FAILURE to compute the
failure count by day and process */
proc sort data=sashelp.failure out=temp;
by day process;
run;

proc summary data=temp;

by day process;
var count;
output out=failure sum=count;
run;

/* Determine the daily status for each process:

count at or under 120% of target is NORM, over is HIGH */
%let target=25;
data failure;
length stat $20;
set failure;
if (count <= (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-NORM"));
if (count > (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-HIGH"));
label day="Weekday" count="Failure Count" stat="Status";
run;

/* Define the template */

proc template;
define statgraph failures;
begingraph;

/* Define the norm/high marker symbols */

symbolchar name=norm char='2714'x;
symbolchar name=high char='2718'x;

/* Create legend entries for the norm/high markers */

legenditem type=marker name="norm" /
markerattrs=(symbol=norm size=15pt color=green)
label="Normal Failure Count";
legenditem type=marker name="high" /
markerattrs=(symbol=high size=15pt color=red)
label="High Failure Count";

/* Define an attribute map for the status markers */

discreteattrmap name="statsymmap";
value "A-NORM" / markerattrs=(symbol=norm color=green);
value "B-NORM" / markerattrs=(symbol=norm color=green);
value "A-HIGH" / markerattrs=(symbol=high color=red);
value "B-HIGH" / markerattrs=(symbol=high color=red);
enddiscreteattrmap;
discreteattrvar attrvar=status var=stat attrmap="statsymmap" ;

/* Define the graph */

entrytitle "Daily Failure Report By Process";
entryfootnote "A failure count that is more than 120% of the
target is considered high.";
layout overlay / yaxisopts=(offsetmax=0.1);
1178 Chapter 12 • Custom Marker Definition Statements

barchartparm category=day response=count / name="barchart"

group=process groupdisplay=cluster;
scatterplot x=day y=eval(count+2) / datalabel=count group=status
groupdisplay=cluster markerattrs=(size=15pt);
referenceline y=&target /
lineattrs=(color=lightred pattern=dot) curvelabel="Target";
discretelegend "barchart" "norm" "high" / down=2
order=columnmajor;
endlayout;
endgraph;
end;
run;

proc sgrender data=failure template=failures;

run;

Program Description
Here is the SAS code for this example.

Create the graph data. The data in Sashelp.Failure is the source data. To get the daily
counts, the data is summarized by weekday and process. The Stat column is then added
to indicate whether the count is acceptable or high. Acceptable values are values that do
not exceed 120% of the target value. Values over 120% are considered high. Acceptable
values are A-NORM and B-NORM, and high values are A-HIGH and B-HIGH, where A
and B identify the process.
/* Summarize the data in SASHELP.FAILURE to compute the
failure count by day and process */
proc sort data=sashelp.failure out=temp;
by day process;
run;

proc summary data=temp;

by day process;
var count;
output out=failure sum=count;
run;

/* Determine the daily status for each process:

count at or under 120% of target is NORM, over is HIGH */
%let target=25;
data failure;
length stat $20;
set failure;
if (count <= (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-NORM"));
if (count > (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-HIGH"));
label day="Weekday" count="Failure Count" stat="Status";
run;

Open the template definition.

/* Define the template */
proc template;
define statgraph failures;
Example: SYMBOLCHAR Statement 1179

begingraph;

Define the custom marker symbols. The NAME= option specifies an identifier for the
markers, and the CHAR= option specifies the Unicode value. Unicode value '2714'x
specifies a heavy check mark, and value '2718'x specifies a heavy ballot X.
/* Define the norm/high marker symbols */
symbolchar name=norm char='2714'x;
symbolchar name=high char='2718'x;

Create legend entries for the custom markers. Suboption SYMBOL= in the
MARKERATTRS= option specifies the identifier of the custom markers.
/* Create legend entries for the norm/high markers */
legenditem type=marker name="norm" /
markerattrs=(symbol=norm size=15pt color=green)
label="Normal Failure Count";
legenditem type=marker name="high" /
markerattrs=(symbol=high size=15pt color=red)
label="High Failure Count";

Create a discrete attribute map to map the custom markers to the Stat column
values. Suboption SYMBOL= in the MARKERATTRS= option specifies the identifier
of the custom markers. Suboption COLOR= sets the color of the Unicode marker
character and the associated data label. The DISCRETEATTRVAR statement creates
attribute map variable STATUS.
/* Define an attribute map for the status markers */
discreteattrmap name="statsymmap";
value "A-NORM" / markerattrs=(symbol=norm color=green);
value "B-NORM" / markerattrs=(symbol=norm color=green);
value "A-HIGH" / markerattrs=(symbol=high color=red);
value "B-HIGH" / markerattrs=(symbol=high color=red);
enddiscreteattrmap;
discreteattrvar attrvar=status var=stat attrmap="statsymmap" ;

Define the graph and close the template definition. The SCATTERPLOT statement
GROUP= option references the STATUS attribute map variable. The EVAL statement in
the Y= option positions each character just above the top of its bar. The SIZE= suboption
of the MARKERATTRS= option sets the marker size. The custom markers are included
in the DISCRETELEGEND statement.
/* Define the graph */
entrytitle "Daily Failure Report By Process";
entryfootnote "A failure count that is more than 120% of the
target is considered high.";
layout overlay / yaxisopts=(offsetmax=0.1);
barchartparm category=day response=count / name="barchart"
group=process groupdisplay=cluster;
scatterplot x=day y=eval(count+2) / datalabel=count group=status
groupdisplay=cluster markerattrs=(size=15pt);
referenceline y=&target /
lineattrs=(color=lightred pattern=dot) curvelabel="Target";
discretelegend "barchart" "norm" "high" / down=2
order=columnmajor;
endlayout;
endgraph;
1180 Chapter 12 • Custom Marker Definition Statements

end;
run;

Generate the graph.

proc sgrender data=failure template=failures;
run;

SYMBOLIMAGE Statement
Defines a marker symbol using an image file so that the image can be referenced in other statements.
Note: This statement is valid in the first maintenance release of SAS 9.4 and later
releases.

Syntax
SYMBOLIMAGE NAME=marker-name IMAGE="image-file-spec" </ option(s)>;

Required Arguments
NAME=marker-name
specifies a name for the marker symbol. The name can be used in statements that
support marker symbols.

Interaction If the name matches one of the system-provided symbol names such as
CIRCLE, then the system symbol is replaced by the user-defined
symbol. See “Marker Options” on page 1350 for a list of the system-
provided symbols.

Note Do not enclose the name in quotation marks.

See “Details” on page 1181

IMAGE="image-file-spec"
specifies the name and location of the image file. The supported image types are GIF,
JPEG, and PNG.

Requirements The image file specification must be enclosed in quotation marks.

The image file must be located on the local file system. URL access
is not supported.

Example image="c:\temp\saslogo.gif"

Optional Arguments
HOFFSET=number
specifies a horizontal offset for the marker symbol.

Default 0. The marker symbol is centered on its data point.

Range -0.5–+0.5, where 0.5 represents one-half of the original marker size.
SYMBOLIMAGE Statement 1181

Note A positive offset moves the marker symbol bounding box to the right. A
negative offset moves it to the left.

ROTATE=number
specifies the angle of rotation for the marker symbol in degrees. Positive angles are
measured in the counter-clockwise direction, and negative angles are measured in the
clockwise direction.

Default 0. No rotation is performed.

Note An angle that exceeds 360 degrees in absolute value can be specified.

SCALE=number
specifies a scale factor for the marker symbol as a percentage. The scale factor is
applied to the character's height and width.

Default 1.0 (100%)

Note The outer edges of the image might be clipped by the bounding box when a
large scale factor is specified.

VOFFSET=number
specifies a vertical offset for the marker symbol.

Default 0. The marker symbol is centered on its data point.

Range -0.5–+0.5, where 0.5 represents one-half of the original marker size.

Note A positive offset moves the marker symbol bounding box up. A negative
offset moves it down.

Details
The SYMBOLIMAGE statement defines a custom marker symbol from an image that is
stored in an image file. The image file must exist on the local file system. URL access is
not supported. The supported image formats are GIF, JPG, and PNG. The marker symbol
that is created can be used in the following options:
• the DATASYMBOLS= option in the BEGINGRAPH statement
• the SYMBOL= suboption of the MARKERATTRS= option, which is supported by
the following statements:

DISCRETELEGEND (VALUE statement) SCATTERPLOT

LEGENDITEM SCATTERPLOTMATRIX
LINECHART SERIESPLOT
NEEDLEPLOT STEPPLOT
Symbol specifications are not checked for uniqueness. More than one SYMBOLIMAGE
statement can define the same character. Therefore, you can use SYMBOLIMAGE
statements in IF/ELSE statements.
The following options normally affect the appearance of markers. However, they have
no effect on image marker symbols that are created by the SYMBOLIMAGE statement.
• the BEGINGRAPH statement DATACONTRASTCOLORS= option
• the COLOR= and WEIGHT= suboptions of the MARKERATTRS= option
1182 Chapter 12 • Custom Marker Definition Statements

• the FILLEDOUTLINEDMARKERS= option

The SIZE= and TRANSPARENCY= suboptions of the MARKERATTRS= option do
affect the appearance of markers that are created by the SYMBOLIMAGE statement.
The markers are clipped to the original marker size after rotation, scaling, and offsets are
applied. If clipping occurs, then you can use the SIZE= suboptions of the
MARKERATTRS= option in the plot statement and the SCALE= option in the
SYMBOLIMAGE statement to adjust the size and scaling to eliminate the clipping.

Example: SYMBOLIMAGE Statement

This example shows how to create marker symbols from GIF icon image files. It also
shows you how to use the symbols in a scatter plot and how to add them to a legend. In
this example, a grouped bar chart shows the daily failure count for two processes,
Process A and Process B. A scatter plot is overlaid on the bar chart to show the total
failure count above each bar. Acceptable counts are indicated by a green check mark
icon. High counts are indicated by a caution icon. Finally, a legend is used to describe
the custom symbols. The following figure shows the output.

Example Program
/* Summarize the data in SASHELP.FAILURE to compute the
failure count by day and process */
proc sort data=sashelp.failure out=temp;
by day process;
run;

proc summary data=temp;

by day process;
var count;
Example: SYMBOLIMAGE Statement 1183

output out=failure sum=count;

run;

/* Determine the daily status for each process:

count at or under 120% of target is NORM, over is HIGH */
%let target=25;
data failure;
length stat $20;
set failure;
if (count <= (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-NORM"));
if (count > (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-HIGH"));
label day="Weekday" count="Failure Count" stat="Status";
run;

/* Define the template */

proc template;
define statgraph failures;
begingraph;

/* Define the norm/high marker symbols */

symbolimage name=norm image="C:\temp\check_green.gif" /
voffset=0.5;
symbolimage name=high image="C:\temp\alert_orange.gif" /
voffset=0.5;

/* Create legend entries for the norm/high markers */

legenditem type=marker name="norm" /
markerattrs=(symbol=norm size=16px)
label="Normal Failure Count";
legenditem type=marker name="high" /
markerattrs=(symbol=high size=16px)
label="High Failure Count";

/* Define an attribute map for the status markers

Note: COLOR= affects marker labels only in this case. */
discreteattrmap name="statsymmap";
value "A-NORM" / markerattrs=(symbol=norm color=green);
value "B-NORM" / markerattrs=(symbol=norm color=green);
value "A-HIGH" / markerattrs=(symbol=high color=orange);
value "B-HIGH" / markerattrs=(symbol=high color=orange);
enddiscreteattrmap;
discreteattrvar attrvar=status var=stat attrmap="statsymmap";

/* Define the graph */

entrytitle "Daily Failure Report By Process";
entryfootnote "A failure count that is more than 120% of the
target is considered high.";
layout overlay / yaxisopts=(offsetmax=0.1);
barchartparm category=day response=count / name="barchart"
group=process groupdisplay=cluster;
scatterplot x=day y=count / datalabel=count group=status
groupdisplay=cluster markerattrs=(size=16px);
referenceline y=&target /
lineattrs=(color=lightred pattern=dot) curvelabel="Target";
discretelegend "barchart" "norm" "high" / down=2
order=columnmajor;
1184 Chapter 12 • Custom Marker Definition Statements

endlayout;
endgraph;
end;
run;

proc sgrender data=failure template=failures;

run;

Program Description
Here is the SAS code for this example.

Create the graph data. The data in Sashelp.Failure is the source data. To get the daily
counts, the data is summarized by weekday and process. The Stat column is then added
to indicate whether the count is acceptable or high. Acceptable values are values that do
not exceed 120% of the target value. Values over 120% are considered high. Acceptable
values are A-NORM and B-NORM, and high values are A-HIGH and B-HIGH, where A
and B identify the process.
/* Summarize the data in SASHELP.FAILURE to compute the
failure count by day and process */
proc sort data=sashelp.failure out=temp;
by day process;
run;

proc summary data=temp;

by day process;
var count;
output out=failure sum=count;
run;

/* Determine the daily status for each process:

count at or under 120% of target is NORM, over is HIGH */
%let target=25;
data failure;
length stat $20;
set failure;
if (count <= (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-NORM"));
if (count > (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-HIGH"));
label day="Weekday" count="Failure Count" stat="Status";
run;

Open the template definition.

/* Define the template */
proc template;
define statgraph failures;
begingraph;

Define the custom marker symbols. The NAME= option specifies an identifier for the
markers, and the IMAGE= option specifies the image file. The VOFFSET= option raises
the markers 50% from their data points in order to prevent them from overlapping the
bars in the bar chart.
/* Define the norm/high marker symbols */
symbolimage name=norm image="C:\temp\check_green.gif" /
Example: SYMBOLIMAGE Statement 1185

voffset=0.5;
symbolimage name=high image="C:\temp\alert_orange.gif" /
voffset=0.5;

Create legend entries for the custom markers. Suboption SYMBOL= in the
MARKERATTRS= option specifies the identifier of the custom markers.
/* Create legend entries for the norm/high markers */
legenditem type=marker name="norm" /
markerattrs=(symbol=norm size=16px)
label="Normal Failure Count";
legenditem type=marker name="high" /
markerattrs=(symbol=high size=16px)
label="High Failure Count";

Create a discrete attribute map to map the custom markers to the Stat column
values. Suboption SYMBOL= in the MARKERATTRS= option specifies the identifier
of the custom markers. Suboption COLOR= sets the color of the associated data label. It
does not affect the color of the image. The DISCRETEATTRVAR statement creates
attribute map variable STATUS.
/* Define an attribute map for the status markers
Note: COLOR= affects marker labels only in this case. */
discreteattrmap name="statsymmap";
value "A-NORM" / markerattrs=(symbol=norm color=green);
value "B-NORM" / markerattrs=(symbol=norm color=green);
value "A-HIGH" / markerattrs=(symbol=high color=orange);
value "B-HIGH" / markerattrs=(symbol=high color=orange);
enddiscreteattrmap;
discreteattrvar attrvar=status var=stat attrmap="statsymmap";

Define the graph and close the template definition. The SCATTERPLOT statement
GROUP= option references the STATUS attribute map variable. The SIZE= suboption
of the MARKERATTRS= option sets the marker size. The custom markers are included
in the DISCRETELEGEND statement.
/* Define the graph */
entrytitle "Daily Failure Report By Process";
entryfootnote "A failure count that is more than 120% of the
target is considered high.";
layout overlay / yaxisopts=(offsetmax=0.1);
barchartparm category=day response=count / name="barchart"
group=process groupdisplay=cluster;
scatterplot x=day y=count / datalabel=count group=status
groupdisplay=cluster markerattrs=(size=16px);
referenceline y=&target /
lineattrs=(color=lightred pattern=dot) curvelabel="Target";
discretelegend "barchart" "norm" "high" / down=2
order=columnmajor;
endlayout;
endgraph;
end;
run;

Generate the graph.

proc sgrender data=failure template=failures;
run;
1186 Chapter 12 • Custom Marker Definition Statements
1187

Part 9

Draw Statements

Chapter 13
Key Concepts for Using Draw Statements . . . . . . . . . . . . . . . . . . . . . . 1189

Chapter 14
Draw Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1199
1188
1189

Chapter 13
Key Concepts for Using Draw
Statements

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1189
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1189
Types of Elements That Can Be Drawn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1192
About the Drawing Space and Drawing Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1192
How the Drawn Elements Are Anchored . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195
About Drawing Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195

Introduction
The draw statements enable you to customize a graph by drawing visual elements
anywhere within the graph. The following sections provide a comprehensive example
and a brief overview to the major drawing concepts. For more details about the
individual draw statements and a simple example of each, see the discussion for each
individual statement. For complete usage information for the draw statements, consult
the SAS Graph Template Language: User's Guide.

Example
The following example shows a bar chart of the top global automobile makers in 2008.
To focus the graph on the 2008 merger between Chrysler and Fiat, the example uses
DRAWRECTANGLE to highlight the three bars in the chart that correspond to the unit
sales for the two auto makers. The example also uses a DRAWARROW statement and
two BEGINPOLYLINE blocks to point to the sales figures, and it uses a DRAWTEXT
statement to explain the implications of the merger.
All of the draw statements in this example use the data space (see “About the Drawing
Space and Drawing Units” on page 1192 ) to integrate the drawn elements into the
graph’s data area.
1190 Chapter 13 • Key Concepts for Using Draw Statements

/* Create the data and the macro variables for Chrysler, Fiat, and
Chrysler+Fiat widths */
data mydata;
length automaker $30;
input automaker $ 1-30 million_units;
FORMAT million_units 3.1;

select (automaker);
when ("Fiat") do;
colorvar=1;
call symput( "fwidth", million_units);
end;
when ("Chrysler") do;
colorvar=1;
call symput( "cwidth", million_units);
end;
when ("Fiat + Chrysler") do;
colorvar=2;
call symput( "cfwidth", million_units);
end;
otherwise colorvar=3;
end;

datalines;
Toyota 8.7
Example 1191

GM 7.7
Volkswagen 6.0
Renault-Nissan 5.8
Ford 5.4
Fiat + Chrysler 4.5
Hyundai 4.2
Honda 3.8
PSA 3.2
Fiat 2.5
Suzuki 2.4
Chrysler 2.0
Daimler 1.9
BMW 1.4
Mazda 1.4
Mitsubishi 1.1
;

/* Create template definition */

proc template;
define statgraph automerger;
begingraph / drawspace=datavalue;
entrytitle halign=center
'Top Global Automakers (2008 Annual Unit Sales)';
layout lattice / rowdatarange=data columndatarange=data
rowgutter=10 columngutter=10;
layout overlay / xaxisopts=(label=('Units (millions)'))
yaxisopts=(reverse=true display=(ticks tickvalues line));
barchart category=automaker response=million_units / group=colorvar
name='bar(h)' barlabel=true dataskin=pressed orient=horizontal;
drawrectangle x=eval(&cwidth/2.0) y="Chrysler"
width=&cwidth height=0.95 / widthunit=data heightunit=data
display=(outline) outlineattrs=(color=yellow);
drawrectangle x=eval(&fwidth/2.0) y="Fiat"
width=&fwidth height=0.95 /
widthunit=data heightunit=data
display=(outline) outlineattrs=(color=yellow);
drawrectangle x=eval(&cfwidth/2.0) y="Fiat + Chrysler"
width=&cfwidth height=0.95 / widthunit=data heightunit=data
display=(outline) outlineattrs=(color=yellow);
beginpolyline x=eval(&cwidth + 0.5) y="Chrysler";
draw x=eval(&cwidth + 1.5) y="Chrysler";
draw x=eval(&cwidth + 1.5) y="Fiat";
draw x=eval(&fwidth + 0.5) y="Fiat";
endpolyline;
beginpolyline x=eval(&cwidth + 1.5) y="Suzuki";
draw x=eval(&cfwidth + 1.5) y="Suzuki";
draw x=eval(&cfwidth + 1.5) y="Fiat + Chrysler";
endpolyline;
drawarrow x1=eval(&cfwidth + 1.5) x2=eval(&cfwidth + 0.5)
y1="Fiat + Chrysler" y2="Fiat + Chrysler"
/ arrowheadscale=0.5 arrowheadshape=barbed;
drawtext "Alliance creates the #6 Global Automaker by volume" /
y="Honda" x=eval(&cfwidth+2.5) width=2 widthunit=data;
endlayout;
endlayout;
endgraph;
1192 Chapter 13 • Key Concepts for Using Draw Statements

end;

proc sgrender data=mydata template=automerger;

run;

Types of Elements That Can Be Drawn

The GTL draw statements enable you to draw the following types of elements in a
graph:
• text
• arrows and lines
• geometric shapes like ovals, rectangles, polygons, and polylines (a set of connecting
lines)
• images
Using these elements individually and especially by combining them, you can describe
the non-data aspects of your graph. For example, you can display a company logo in a
specified location within the graph. Or you can create custom features that are difficult
to create by other means—for example, you could draw a broken axis. Using the draw
statements creatively, you can direct viewer attention to features of interest in the graph
by drawing elements that highlight those features.

About the Drawing Space and Drawing Units

Each draw statement positions a drawn element using Cartesian coordinates that you
specify as X and Y values in the statement. The coordinates that you specify are
positioned relative to the drawing space that is in effect for the statement. The available
drawing spaces are the data area, the wall area, the layout area, and the graph area. The
coordinates can be specified in pixel units, percentage units, or as values that are in the
units of the data (available only in the data area).
The images that accompany the following descriptions all use pixel data units. Each,
however, was drawn in a different data space using this same LAYOUT OVERLAY
definition block:
layout overlay /
xaxisopts=(display=(line ticks tickvalues))
yaxisopts=(display=(line ticks tickvalues));
scatterplot x=height y=weight;
drawrectangle x=0 y=0 width=50 height=50 /
anchor=bottomleft display=(fill) fillattrs=(color=green)
transparency=0.75 widthunit=pixel heightunit=pixel ;
drawline x1=0 y1=0 x2=18 y2=18 / lineattrs=(color=red) ;
endlayout;

Drawing Space: Data Area

Drawing Units: VALUE, PERCENT, PIXEL

About the Drawing Space and Drawing Units 1193

Drawing Space: Data Area

The area where data is displayed in the graph,

honoring the offsets that are set for the axes. For a
discussion on axis offsets, see “Adjusting Axis
Offsets” on page 886.
The data area does not apply to graphs that do not
have axes, such as pie charts, which must be drawn
in a REGION layout.
For graphs produced within LAYOUTDATAPANEL
and LAYOUTDATALATTICE layouts, drawn
elements are clipped if they extend outside of the
wall boundaries.

Drawing Space: Wall Area

Drawing Units: PERCENT, PIXEL

The area bounded by orthogonal axis pairs, ignoring

the offsets that are set for the axes. In two-
dimensional graphs, there is one wall bounded by the
XY axes. including the secondary axes, if used. In
three-dimensional graphs, there are three walls,
bounded by the XY, YZ, and XZ axes.
The wall area does not apply to graphs that do not
have axes, such as pie charts, which must be drawn
in a REGION layout.
For graphs produced within LAYOUTDATAPANEL
and LAYOUTDATALATTICE layouts, drawn
elements that extend outside of the wall boundaries
are clipped.

Drawing Space: Layout Area

Drawing Units: PERCENT, PIXEL

The entire area of the layout container that is the

immediate parent container of the draw statement.
The figure to the right shows the case where a
LAYOUT OVERLAY is the draw statement’s layout
container.
Titles and footnotes are always displayed outside of
the outermost layout, so those areas are never part of
the layout drawing space.

Drawing Space: Graph Area

Drawing Units: PERCENT, PIXEL

1194 Chapter 13 • Key Concepts for Using Draw Statements

Drawing Space: Graph Area

The entire area that is available to the graph display,

whether a single-cell or multi-cell graph.
Because the graph drawing space spans the entire
graph, the location of the drawn element in the graph
is independent of the draw statement’s placement
within the template definition, even if the draw
statement is specified within a nested layout.

To specify the drawing space for both the X and the Y dimension, you use the
DRAWSPACE= option. To specify the drawing space individually for either the X or the
Y dimension, you use the options XSPACE=, X1SPACE=, X2SPACE=, YSPACE=,
Y1SPACE=, or Y2SPACE=, depending on the draw statement. The value that you set on
any of these options is a single composite value that specifies both the drawing space
and the drawing units in the following format:
<DrawingSpace><Units>

For example, DRAWSPACE=GRAPHPIXEL specifies the GRAPH drawing space with

PIXEL drawing units, indicating that the statement’s X,Y coordinates are expressed in
pixels. Similarly, DRAWSPACE=LAYOUTPERCENT specifies the LAYOUT drawing
space with PERCENT drawing units, indicating that the statement’s X,Y coordinates are
expressed as percentages.
The global DRAWSPACE is LAYOUTPERCENT. The global DRAWSPACE= setting
for all of the draw statements is LAYOUTPERCENT. The draw statements inherit the
global setting from the DRAWSPACE= setting in the BEGINGRAPH statement.
• To change the global drawing space and drawing units for all of the draw statements
within the template definition, use the DRAWSPACE= option in the BEGINGRAPH
statement.
• To change the default for an individual draw statement, use that statement’s
DRAWSPACE= option. If needed, you can specify different settings for the different
draw statements within the template definition.
For the DATA drawing space, the VALUE drawing units specify that the coordinates are
expressed as values along the axis. When you specify the DATA drawing space, you can
use the draw statement’s XAXIS= and YAXIS= options to specify which axis scale to
use for the coordinates.
• If the specified axis does not exist in the plot or is not valid for the draw statement's
layout container, then the draw statement is ignored.
• For a discrete axis, if the statement’s specified X or Y value does not exist in the
data, then the draw statement is ignored.
• For a continuous axis, if the statement’s specified X or Y value does not exist in the
data, then the value is extrapolated.
When specifying the drawing space and drawing units, you can set a common setting for
all of the X and Y coordinates. Or you can specify different settings for each individual
coordinate. The DRAWSPACE= setting in the BEGINGRAPH statement applies the
global space and unit settings to all of the draw statements within the BEGINGRAPH/
ENDGRAPH block. The DRAWSPACE= setting in an individual draw statement applies
About Drawing Layers 1195

the space and unit setting only to the coordinate(s) for that statement. Thus, for lines and
arrows, the setting applies to both the X1, Y1 coordinate and the X2, Y2 coordinate.
To specify the drawing space and drawing units separately for the X coordinate and for
the Y coordinate, use the XSPACE=, YSPACE=, X1SPACE=, Y1SPACE=, X2SPACE=,
and Y2SPACE= options, as applicable, in each draw statement. These options override
the DRAWSPACE= option.

How the Drawn Elements Are Anchored

When you specify the X and Y coordinates in a draw statement, the element is drawn
from an anchor point that is placed in the drawing area at the X,Y coordinate point.
• For lines and arrows, the anchor point is the starting point of the line or arrow, which
is specified with the draw statement’s X1= and Y1= values.
• For elements that have height and width, the anchor point can be one of the
following points:

The default anchor position is CENTER. To change the anchor position, use the draw
statement’s ANCHOR= option.

About Drawing Layers

A draw statement can draw its element in either of two “layers” in the graph: the front
layer or the back layer. By default, the statement draws in the front layer, which places
the element in front of all other graphics elements, including data points, data labels, axis
labels, and so on. In some cases, this might cause the drawn element to block the view of
other graphics elements in the graph.
To prevent a drawn element for blocking the view of other graphics elements in the
graph, you can use the draw statement’s TRANSPARENCY= option to add transparency
to the drawn element. With an appropriate transparency setting, you should be able to
see any graphics elements that are behind the drawn element.
Alternatively, you can use the draw statement’s LAYER= option to draw the element in
the back layer, which places the drawn element behind all other graphics elements in the
graph. For example, the following figure shows two different versions of a graph that
uses DRAWLINE to draw a diagonal line across the axis area. The version to the left
draws the line in the front layer, which causes the line to cover some of the data labels in
the graph and portions of the series line. The version to the right uses LAYER=BACK to
draw the line in the back layer. This prevents the line from covering the data labels and
the portions of the series line.
1196 Chapter 13 • Key Concepts for Using Draw Statements

The following code fragment shows the code that positions the line in the back layer:
layout overlay / walldisplay=(outline)
xaxisopts=(griddisplay=on display=(line ticks tickvalues))
yaxisopts=(griddisplay=on display=(line ticks tickvalues));
seriesplot x=open y=close / datalabel=date;
drawline x1=0 y1=0 x2=100 y2=100 /
x1space=wallpercent y1space=wallpercent
x2space=wallpercent y2space=wallpercent
lineattrs=(color=cyan thickness=6) layer=back ;
endlayout;

The Back Layer is Behind the Background. Although drawing elements in the back
layer prevents them from obstructing other data elements in the graph, it is not always
the right solution to the problem.
If a draw statement uses LAYER=BACK, then it draws the element behind all other
graphics elements, such as the layout background or a discrete legend’s background. To
ensure that the element is visible in the graph, you might have to do one or more of the
following:
• In overlay-type layouts or in a SCATTERPLOTMATRIX, you can use the parent
layout’s WALLDISPLAY= option to turn off the display of the plot wall. In the
example code just shown, WALLDISPLAY=(OUTLINE) displays an outline around
the graph wall but does not display the wall fill. Suppressing the fill ensures that the
drawn line is visible behind the plot wall.
• If a layout container uses OPAQUE=TRUE so that it can set visual attributes for the
background, then the opaque background covers and therefore hides any drawn
element that is behind the background. When assigning visual attributes to a graph
background, therefore, it might be better to use TRANSPARENCY= rather than
LAYER= to prevent drawn elements from covering other graphics elements in the
graph.
• If the results of a plot statement or other GTL-statement covers the drawn element,
then you can use transparency to reveal the drawn element. For example, you could
About Drawing Layers 1197

use a plot statement’s DATATRANSPARENCY= option to set an appropriate

transparency level for the plot.
1198 Chapter 13 • Key Concepts for Using Draw Statements
1199

Chapter 14
Draw Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1199
BEGINPOLYGON Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1199
BEGINPOLYLINE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1206
DRAWARROW Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1212
DRAWIMAGE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1219
DRAWLINE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1227
DRAWOVAL Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1233
DRAWRECTANGLE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1241
DRAWTEXT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1249

Dictionary

BEGINPOLYGON Statement
Defines the starting point for drawing a polygon in the graph.

Syntax
BEGINPOLYGON X=constant | scalar-expression
Y=constant | scalar-expression </option(s)>;
DRAW X=constant | scalar-expression
Y=constant | scalar-expression </option(s)>;
<… more DRAW statements …>
ENDPOLYGON;

Summary of Optional Arguments

Appearance options
DISPLAY=STANDARD | ALL | (display-options)
specifies the features to display for the polygon.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the fill attributes for this BEGINPOLYGON block.
LAYER=FRONT | BACK
specifies the layer on which this BEGINPOLYGON block’s output is drawn.
1200 Chapter 14 • Draw Statements

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the outline attributes for this BEGINPOLYGON block.
TRANSPARENCY=number
specifies the degree of the transparency of this BEGINPOLYGON block’s
output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted
using the primary X axis scale or to the secondary X (X2) axis scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted
using the primary Y axis scale or to the secondary Y (Y2) axis scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this
BEGINPOLYGON block.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the Y= option.

ODS options
URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Required Arguments
The following options are required in the BEGINPOLYGON statement and the DRAW
statement.
X=constant | scalar-expression
specifies the X value for a point in the polygon. When used in the
BEGINPOLYGON statement, it specifies the X value of the starting point of the
polygon.

Interaction The value that is set for this argument is interpreted using the
XSPACE= option. When XSPACE=DATAVALUE, the value is
interpreted using the XAXIS= option.
BEGINPOLYGON Statement 1201

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y=constant | scalar-expression
specifies the Y value for a point in the polygon. When used in the
BEGINPOLYGON statement, it specifies the Y value of the starting point of the
polygon.

Interaction The value that is set for this argument is interpreted using the
YSPACE= option. When YSPACE=DATAVALUE, the value is
interpreted using the YAXIS= option.

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Optional Arguments
The following options can be used in the BEGINPOLYGON statement.
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y values, or
both.

Default 0 (no offset, output is centered on the discrete X values, or the discrete
Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If the
Y axis is discrete, then a positive offset is up. If REVERSE=TRUE on
the X or Y axis, then the offset direction is also reversed.

Restriction This option applies only when the options XSPACE= or YSPACE= use
DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DISPLAY=STANDARD | ALL | (display-options)

specifies the features to display for the polygon.
STANDARD
displays an outlined polygon.
ALL
displays an outlined, filled polygon.
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays an outlined polygon

FILL displays a filled polygon

1202 Chapter 14 • Draw Statements

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the polygon.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this BEGINPOLYGON
block.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting for
only this BEGINPOLYGON statement.

This option sets the default drawing space, but individual settings in
the X or Y dimension can be overridden by the XSPACE= and
YSPACE= options.

See “About the Drawing Space and Drawing Units” on page 1192

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the fill attributes for this BEGINPOLYGON block.

Default The GraphAnnoShape style element

Tip The TRANSPARENCY= option sets the transparency for the fill and the
outline. You can combine this option with TRANSPARENCY= to set one
transparency for the outlines but a different transparency for the fill.
Example:
transparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

LAYER=FRONT | BACK
specifies the layer on which this BEGINPOLYGON block’s output is drawn.

FRONT draws the output on top of the graph.

BACK draws the output behind the background areas, such as a layout or
legend background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer, you can
suppress the display of filled areas in the graph. You can also use
transparency to manage the element visibility. For more information, see
“About Drawing Layers” on page 1195 .

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the outline attributes for this BEGINPOLYGON block.
BEGINPOLYGON Statement 1203

Default The GraphAnnoShape style element.

Interaction For this option to have any effect, the outline must be enabled by the
DISPLAY= option.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element .

“Line Options” on page 1349 for available line-options.

TRANSPARENCY=number
specifies the degree of the transparency of this BEGINPOLYGON block’s output.

Default 0

Range 0 (opaque) to 1 (entirely transparent)

Tip The FILLATTRS option can be used to set transparency for just the
polygon’s filled area. You can combine this option with FILLATTRS= to
set one transparency for the outlines but a different transparency for the fill.
Example:
transparency=0.2 fillattrs=(transparency=0.6)

URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw statement is
selected.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
string | string-expression
specifies a valid HTML page reference (HREF) for the graphical element that is
drawn by this draw statement.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable graphical elements, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted using
the primary X axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

1204 Chapter 14 • Draw Statements

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted using
the primary Y axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

DRAW Statement Optional Arguments

The following options can be used in the DRAW statement.

Statement Option Description

DRAWSPACE Specifies a default drawing space and drawing

units for the drawn lines.

XSPACE Specifies the drawing space and drawing units

for interpreting the X value.

YSPACE Specifies the drawing space and drawing units

for interpreting the Y value.

Details

Statement Description
A polygon is built by using the BEGINPOLYGON statement to specify the polygon’s
starting X,Y coordinate, and then specifying the remaining points by nesting a series of
DRAW statements (see “DRAW Statement” on page 1204 ) within the
BEGINPOLYGON block. The block is closed with an ENDPOLYGON statement. To
manage the location and drawing units for the polygon, you can use the XAXIS=,
YAXIS=, XSPACE=, and YSPACE= options.
For general information about the types of elements that can be drawn with the draw
statements, the drawing space and drawing units that they use, and how the drawn
elements are anchored, see Chapter 13, “Key Concepts for Using Draw Statements,” on
page 1189 . For detailed usage information, consult the SAS Graph Template Language:
User's Guide.

DRAW Statement
The nested DRAW statements within a BEGINPOLYGON block specify a series of
points for a polygon. Each DRAW statement draws a straight line from the previous
point to the endpoint that is specified in the DRAW statement's X and Y arguments. The
first DRAW statement starts its line from the X,Y point that is specified in the
BEGINPOLYGON statement.
Example: BEGINPOLYGON Statement 1205

You can specify as many DRAW statements as needed to complete the polygon. If the
last DRAW statement does not end at the first point in the polygon (specified in
BEGINPOLYGON), then a line is automatically generated to close the polygon shape.
For a specific example, see the “Example Program” on page 1205.

Example: BEGINPOLYGON Statement

The following graph was generated by the “Example Program” on page 1205. The
example uses a BEGINPOLYGON block to draw a polygon around the Setosa species of
Iris in a plot that shows petal sizes for three Iris species. The DRAWSPACE= option in
the BEGINGRAPH statements specifies that the polygon be drawn in the data space.
The BEGINPOLYGON statement specifies the starting X,Y point. For the
BEGINPOLYGON options, DISPLAY= displays only the fill for the polygon.
TRANSPARENCY= adds a degree of transparency to the fill, and FILLATTRS= sets the
fill color to yellow. The example also uses DRAWARROW and DRAWTEXT
statements to draw an annotation for the polygon.

Example Program
proc template;
define statgraph discretelegend;
begingraph / drawspace=datavalue;
entrytitle "Iris Petal Dimensions";
layout overlayequated / equatetype=equate;
scatterplot x=petallength y=petalwidth / group=species name="s";
ellipse x=petallength y=petalwidth / type=predicted alpha=.2
name="p80" legendlabel="80%" outlineattrs=graphconfidence;
ellipse x=petallength y=petalwidth / type=predicted alpha=.05
name="p95" legendlabel="95%" outlineattrs=graphconfidence2;
1206 Chapter 14 • Draw Statements

beginpolygon x=9 y=2 /display=(fill)

fillAttrs=(color=yellow transparency=0.75);
draw x=13 y=5;
draw x=16 y=7;
draw x=17 y=6;
draw x=20 y=5;
draw x=20 y=1;
draw x=17 y=1;
draw x=15 y=0;
draw x=14 y=0;
draw x=11 y=0;
endpolygon;
drawtext textattrs=(size=11pt) "Setosa" /
x=33 y=0 width=15 anchor=top;
drawarrow x1=28 y1=-1 x2=19 y2=2 /
lineattrs=(thickness=1px)
arrowheadshape=barbed arrowheadscale=0.1;
discretelegend "s" / title="Species:";
discretelegend "p80" "p95" /across=1 autoalign=(topleft) location=inside;
endlayout;
entryfootnote halign=left "Fisher's Iris Data" ;
endgraph;
end;
run;

proc sgrender data=sashelp.iris template=discretelegend;

run;

BEGINPOLYLINE Statement
Defines the starting point for drawing a polyline in the graph.

Syntax
BEGINPOLYLINE X=constant | scalar-expression
Y=constant | scalar-expression </option(s)>;
DRAW X=constant | scalar-expression
Y=constant | scalar-expression </option(s)>;
<… more DRAW statements …>
ENDPOLYLINE;

Summary of Optional Arguments

Appearance options
LAYER=FRONT | BACK
specifies the layer on which this BEGINPOLYLINE block’s output is drawn.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line attributes for this BEGINPOLYLINE block.
TRANSPARENCY=number
specifies the degree of the transparency of the BEGINPOLYLINE block’s
output.
BEGINPOLYLINE Statement 1207

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted
using the primary X axis scale or to the secondary X (X2) axis scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted
using the primary Y axis scale or to the secondary Y (Y2) axis scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this
BEGINPOLYLINE block.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the Y= option.

ODS options
URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Required Arguments
The following options are required in the BEGINPOLYLINE statement and the DRAW
statement.
X=constant | scalar-expression
specifies the X value for the starting point in the polyline.

Interaction The value that is set for this argument is interpreted using the
XSPACE= option. When XSPACE=DATAVALUE, the value is
interpreted using the XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y=constant | scalar-expression
specifies the Y value for the starting point in the polyline.

Interaction The value that is set for this argument is interpreted using the
YSPACE= option. When YSPACE=DATAVALUE, the value is
interpreted using the YAXIS= option.
1208 Chapter 14 • Draw Statements

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Optional Arguments
The following options can be used in the BEGINPOLYLINE statement.
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y values, or
both.

Default 0 (no offset, output is centered on the discrete X values, or the discrete
Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If the
Y axis is discrete, then a positive offset is up. If REVERSE=TRUE on
the X or Y axis, then the offset direction is also reversed.

Restriction This option applies only when the options XSPACE= or YSPACE= use
DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this BEGINPOLYLINE
block.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting for
only this BEGINPOLYLINE statement.

This option sets the default drawing space, but individual settings in
the X or Y dimension can be overridden by the XSPACE= and
YSPACE= options.

See “About the Drawing Space and Drawing Units” on page 1192

LAYER=FRONT | BACK
specifies the layer on which this BEGINPOLYLINE block’s output is drawn.

FRONT draws the output on top of the graph.

BACK draws the output behind the background areas, such as a layout or
legend background.
BEGINPOLYLINE Statement 1209

Default FRONT

Tip For elements that are obstructed because they are in the back layer, you can
suppress the display of filled areas in the graph. You can also use
transparency to manage the element visibility. For more information, see
“About Drawing Layers” on page 1195 .

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes for this BEGINPOLYLINE block.

Default The GraphAnnoLine style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element .

“Line Options” on page 1349 for available line-options.

TRANSPARENCY=number
specifies the degree of the transparency of the BEGINPOLYLINE block’s output.

Default 0

Range 0 (opaque) to 1 (entirely transparent)

URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw statement is
selected.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
string | string-expression
specifies a valid HTML page reference (HREF) for the graphical element that is
drawn by this draw statement.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable graphical elements, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted using
the primary X axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

1210 Chapter 14 • Draw Statements

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted using
the primary Y axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

DRAW Statement Optional Arguments

The following options can be used in the DRAW statement.

Statement Option Description

DRAWSPACE Specifies a default drawing space and drawing

units for the drawn lines.

XSPACE Specifies the drawing space and drawing units

for interpreting the X value.

YSPACE Specifies the drawing space and drawing units

for interpreting the Y value.

Details

Statement Description
A polyline is built by using the BEGINPOLYLINE statement to specify the polyline’s
starting X,Y coordinate, and then specifying the remaining points by nesting a series of
DRAW statements (see “DRAW Statement” on page 1210 ) within the
BEGINPOLYLINE block. The block is closed with an ENDPOLYLINE statement. To
manage the location and drawing units for the polyline, you can use the XAXIS=,
YAXIS=, XSPACE=, and YSPACE= options.
For general information about the types of elements that can be drawn with the draw
statements, the drawing space and drawing units that they use, and how the drawn
elements are anchored, see Chapter 13, “Key Concepts for Using Draw Statements,” on
page 1189 . For detailed usage information, consult the SAS Graph Template Language:
User's Guide.

DRAW Statement
The nested DRAW statements within a BEGINPOLYLINE block specify a series of
points for a polyline. Each DRAW statement draws a straight line from the previous
point to the endpoint that is specified in the DRAW statement's X and Y arguments. The
first DRAW statement starts its line from the X,Y point that is specified in the
BEGINPOLYLINE statement. You can specify as many DRAW statements as needed to
complete the polyline.
Example: BEGINPOLYLINE Statement 1211

For a specific example, see the “Example Program” on page 1211.

Example: BEGINPOLYLINE Statement

The following graph was generated by the “Example Program” on page 1211. The
example uses two BEGINPOLYLINE blocks to highlight two student data points. The
DRAWSPACE= option in the BEGINGRAPH statements specifies that the polyline be
drawn in the data space. The BEGINPOLYLINE statements specify the starting X,Y
points for two polyline, and the DRAW statements complete the lines. The example also
uses two DRAWTEXT statements to label the data points of interest.

Example Program
proc template;
define statgraph drawoval;
begingraph / drawspace=datavalue;
entrytitle "Regression Fit Plot";
layout overlay;
modelband "myclm";
scatterplot x=height y=weight;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
drawtext "Alfred" / x=69 y=112 anchor=top;
drawtext "Barbara" / x=65.4 y=97 anchor=top width=15;
beginpolyline x=69 y=105 ;
draw x=69 y=85 ;
draw x=65.3 y=85 ;
draw x=65.3 y=90 ;
endpolyline ;
beginpolyline x=67 y=85 ;
1212 Chapter 14 • Draw Statements

draw x=67 y=78 ;

endpolyline ;
drawtext "New in Class" / x=67 y=77 anchor=top width=15;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=drawoval;

run;

DRAWARROW Statement
Draws an arrow (a directed line segment) from one point to another point.

Syntax
DRAWARROW X1=constant | scalar-expression
Y1=constant | scalar-expression
X2=constant | scalar-expression
Y2=constant | scalar-expression </option(s)>;

Summary of Optional Arguments

Appearance options
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line attributes for this draw statement.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Arrow options
ARROWHEADDIRECTION=OUT | IN | BOTH
specifies the direction of the arrowhead(s) at the end(s) of the arrow shaft.
ARROWHEADSCALE=positive-number
specifies an arrowhead scale factor based on the thickness of the arrow shaft.
ARROWHEADSHAPE=OPEN | CLOSED | FILLED | BARBED
specifies the shape of the arrowhead(s).

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted
using the primary X axis scale or to the secondary X (X2) axis scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted
using the primary Y axis scale or to the secondary Y (Y2) axis scale.

Drawing space options

DRAWARROW Statement 1213

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.
X1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X1 value.
X2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X2 value
Y1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y1 value.
Y2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y2 value

ODS options
URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Required Arguments
X1=constant | scalar-expression
specifies the X value of one arrow-shaft endpoint.

Interactions This value that is set for this option is interpreted using the
X1SPACE= option.

When X1SPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When X1SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y1=constant | scalar-expression
specifies the Y value of one arrow-shaft endpoint.

Interactions This value that is set for this option is interpreted using the
Y1SPACE= option.

When Y1SPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When Y1SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

X2=constant | scalar-expression
specifies the X value of one arrow-shaft endpoint.
1214 Chapter 14 • Draw Statements

Interactions This value that is set for this option is interpreted using the
X2SPACE= option.

When X2SPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When X2SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y2=constant | scalar-expression
specifies the Y value of one arrow-shaft endpoint.

Interactions This value that is set for this option is interpreted using the
Y2SPACE= option.

When Y2SPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When Y2SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Optional Arguments
ARROWHEADDIRECTION=OUT | IN | BOTH
specifies the direction of the arrowhead(s) at the end(s) of the arrow shaft.

OUT specifies a single arrowhead drawn at (X2,Y2) and pointing away from
(X1,Y1)
IN specifies a single arrowhead drawn at (X1,Y1) and pointing away from
(X2,Y2)
BOTH specifies two arrowheads, one at the IN position and one at the OUT
position

Default OUT

Tip Use the ARROWHEADSHAPE= option to control the arrowhead

appearance.

ARROWHEADSHAPE=OPEN | CLOSED | FILLED | BARBED

specifies the shape of the arrowhead(s). The following figure shows the arrowhead
shapes.

Default OPEN
DRAWARROW Statement 1215

Tip Use the ARROWHEADDIRECTION= option to control the arrow

direction.

ARROWHEADSCALE=positive-number
specifies an arrowhead scale factor based on the thickness of the arrow shaft.

Default 1.0

Restriction The minimum size for arrowheads is 8 pixels. If you specify a value for
ARROWHEADSCALE= that scales the arrowhead below 8 pixels, an
8-pixel arrowhead is used instead. There is no restriction on the
maximum size.

Tip Use a factor greater than 1.0 to make a larger arrowhead.

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y values, or
both.

Default 0 (no offset, output is centered on the discrete X values, or the discrete
Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If the
Y axis is discrete, then a positive offset is up. If REVERSE=TRUE on
the X or Y axis, then the offset direction is also reversed.

Restriction This option applies only when the options X1SPACE=, X2SPACE=,
Y1SPACE=, or Y2SPACE= use DATAVALUE, and when X1, X2, Y1,
or Y2 are values on a discrete axis. For nondiscrete axes, this option is
ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting for
only this draw statement.

This option sets the default drawing space, but individual settings in
the X or Y dimension can be overridden by the options X1SPACE=,
Y1SPACE=, X2SPACE=, or Y2SPACE=.

See “About the Drawing Space and Drawing Units” on page 1192
1216 Chapter 14 • Draw Statements

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT draws the output on top of the graph.

BACK draws the output behind the background areas, such as a layout or
legend background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer, you can
suppress the display of filled areas in the graph. You can also use
transparency to manage the element visibility. For more information, see
“About Drawing Layers” on page 1195 .

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes for this draw statement.

Default The GraphAnnoLine style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw statement is
selected.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
string | string-expression
specifies a valid HTML page reference (HREF) for the graphical element that is
drawn by this draw statement.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable graphical elements, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted using
the primary X axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if X1SPACE=DATAVALUE.

DRAWARROW Statement 1217

X1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X1 value.

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the X1 value.

X2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X2 value

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the X2 value.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted using
the primary Y axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if Y1SPACE=DATAVALUE.

Y1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y1 value.

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the Y1 value.

Y2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y2 value

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the Y2 value.

Details
A DRAWARROW statement draws a line (arrow shaft) from a specified starting point
(X1,Y1) to a specified ending point (X2,Y2). It also displays an arrowhead at either or
both ends of the line. DRAWARROW is similar to a DRAWLINE statement, using many
of the same options, but it has additional options for controlling the arrowhead(s).
For general information about the types of elements that can be drawn with the draw
statements, the drawing space and drawing units that they use, and how the drawn
elements are anchored, see Chapter 13, “Key Concepts for Using Draw Statements,” on
page 1189 . For detailed usage information, consult the SAS Graph Template Language:
User's Guide.
1218 Chapter 14 • Draw Statements

Example: DRAWARROW Statement

The following graph was generated by the “Example Program” on page 1218. The
example shows a common application of a DRAWARROW and DRAWTEXT
statements to identify a specific part of the graph and add explanatory text.

Example Program
proc template;
define statgraph arrow;
begingraph;
entrytitle "Micosoft Stock Prices between 2000 and 2002";
layout overlay;
seriesplot x=date y=close;
drawarrow x1="01NOV2001"d y1=75 x2="01NOV2001"d y2=64.21 /
x1space=datavalue y1space=wallpercent
x2space=datavalue y2space=datavalue
arrowheadshape=filled lineattrs=(color=red) ;
drawtext "Introduction of Windows XP" / width=25 anchor=bottom
border=true borderattrs=(color=red)
x="01NOV2001"d y=75 xspace=datavalue yspace=wallpercent;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.stocks template=arrow;

where stock="Microsoft" and Year(date) between 2000 and 2002;
run;
DRAWIMAGE Statement 1219

DRAWIMAGE Statement
Draws an image in the graph.

Syntax
DRAWIMAGE "image-file-spec" </option(s)>;

Summary of Optional Arguments

Appearance options
BORDER=TRUE | FALSE
specifies whether a border is drawn around the image.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the border line attributes for this draw statement.
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
ROTATE=number
specifies the angle of rotation for the image, measured in degrees.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted
using the primary X axis scale or to the secondary X (X2) axis scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted
using the primary Y axis scale or to the secondary Y (Y2) axis scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the Y= option.
1220 Chapter 14 • Draw Statements

Location options
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the image.
X=constant | scalar-expression
specifies the anchor point’s X coordinate.
Y=constant | scalar-expression
specifies the anchor point’s Y coordinate.

ODS options
URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Size options
HEIGHT=positive-number
specifies the height of the image’s bounding box.
HEIGHTUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the HEIGHT=
option is interpreted as a percentage value, a pixel value, or a value that is in
the units of the data.
SCALE=FIT | FITHEIGHT | FITWIDTH | TILE
specifies how the image is scaled within the bounding box.
SIZEUNIT=PERCENT | PIXEL | DATA
specifies whether the default units for the size of the image’s bounding box
are percentage values, or pixel values, or values that are in the unit of the
data.
WIDTH=positive-number
specifies the width of the image’s bounding box.
WIDTHUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the WIDTH= option
is interpreted as a percentage value, a pixel value, or a value that is in the
units of the data.

Required Argument
image-file-spec
specifies the name, image type, and location of the image. The image-file-spec value
must be enclosed in double quotation marks and must be specified as a local,
physical file path (for example, "c:\temp\saslogo.gif"). The supported image
types are GIF, JPEG, and PNG raster or bitmap format.

Restrictions URL access to image files is not supported. The image file must exist
on the file system.

In SAS 9.4 and in earlier releases, you cannot use a dynamic variable
for the image-file-spec value. This restriction is removed starting with
the first maintenance release of SAS 9.4.
DRAWIMAGE Statement 1221

Optional Arguments
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the image. The anchor point is relative to the unrotated
image. The anchor point can be at the center of the image or at eight points on the
border of the image bounding box. The following figure shows the anchor points for
TOPLEFT and LEFT.

The coordinates of the anchor point are set by the X= and Y= options, and by the
XSPACE= and YSPACE= options. The XAXIS= and YAXIS= option might affect
positioning when the XSPACE= or YSPACE= options are set to DATAPIXEL,
DATAPERCENT, or DATAVALUE.
The image has a fixed height and a fixed width, determined by the HEIGHT=,
HEIGHTUNIT=, WIDTH= and WIDTHUNIT= options. The height of the text
grows in a direction that is related to the anchor point. For example, if
ANCHOR=TOPLEFT, then the image height extends downward from the anchor
point and its width extends to the right. If ANCHOR=CENTER, then half the image
width and half the image height extend equally left and right, as well as top to
bottom from the anchor point. If ANCHOR=BOTTOM, then the image height
extends upward from the anchor point and the image width is centered at the anchor
point.
When the image is rotated, the anchor point remains relative to the unrotated image
while the image is rotated on its anchor point. See ROTATE= on page 1223.

Default CENTER

BORDER=TRUE | FALSE
specifies whether a border is drawn around the image.

Default FALSE

Tip Use the BORDERATTRS= option to control the appearance of the border.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the border line attributes for this draw statement.

Default The GraphBorderLines style element.

Interaction BORDER=TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

1222 Chapter 14 • Draw Statements

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y values, or
both.

Default 0 (no offset, output is centered on the discrete X values, or the discrete
Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If the
Y axis is discrete, then a positive offset is up. If REVERSE=TRUE on
the X or Y axis, then the offset direction is also reversed.

Restriction This option applies only when the options XSPACE= or YSPACE= use
DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting for
only this draw statement.

This option sets the default drawing space, but individual settings in
the X or Y dimension can be overridden by the options XSPACE=,
YSPACE=, HEIGHTUNIT=, or WIDTHUNIT=.

See “About the Drawing Space and Drawing Units” on page 1192

HEIGHT=positive-number
specifies the height of the image’s bounding box.

Default The height of the image.

Interaction The interpretation of this height setting is determined by the combined

settings of the HEIGHTUNIT= and YSPACE= options.

HEIGHTUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the HEIGHT= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of the
data.

Default PERCENT

Interaction This setting combines with the YSPACE= setting to interpret the height
that is set in the HEIGHT= option.
DRAWIMAGE Statement 1223

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT draws the output on top of the graph.

BACK draws the output behind the background areas, such as a layout or
legend background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer, you can
suppress the display of filled areas in the graph. You can also use
transparency to manage the element visibility. For more information, see
“About Drawing Layers” on page 1195 .

ROTATE=number
specifies the angle of rotation for the image, measured in degrees. The image is
rotated around its anchor point. The angle is measured from a horizontal line passing
through the anchor point of the image to the right. The following figure shows the
rotation of an image around a top anchor point.

Positive angles rotate the image counter clockwise, and negative angles rotate the
image clockwise. The angle specification can exceed 360 degrees in absolute value.

Default 0. No rotation is performed.

Note If you specify the SVG format for your graph and your graph includes a
rotated image, then the graph is rendered as a bitmapped image rather than
an SVG image. In that case, a note is written to the SAS log indicating the
change in format.

See ANCHOR= on page 1221

SCALE=FIT | FITHEIGHT | FITWIDTH | TILE

specifies how the image is scaled within the bounding box.
FIT
scales the image to fit the bounding box. Aspect ratio is not maintained.
FITHEIGHT
scales the image to fit the height of the bounding box. The width is computed
from the height and the image's aspect ratio.
1224 Chapter 14 • Draw Statements

FITWIDTH
scales the image to fit the width of the bounding box. The height is computed
from the width and the image's aspect ratio.
TILE
tiles the image as needed to fit the bounding image. The last tile in a row or
column might be clipped by the bounding box.

Default FIT

SIZEUNIT=PERCENT | PIXEL | DATA

specifies whether the default units for the size of the image’s bounding box are
percentage values, or pixel values, or values that are in the unit of the data.

Default PERCENT

Interaction If the HEIGHTUNIT= or WIDTHUNIT= option is also used, then it

overrides this option for that dimension.

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option also affects the image border.

URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw statement is
selected.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
string | string-expression
specifies a valid HTML page reference (HREF) for the graphical element that is
drawn by this draw statement.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable graphical elements, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

WIDTH=positive-number
specifies the width of the image’s bounding box.

Default The width of the image.

WIDTHUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the WIDTH= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of the
data.

Default PERCENT
DRAWIMAGE Statement 1225

Interaction This setting combines with the XSPACE= setting to interpret the width
that is set in the WIDTH= option.

X=constant | scalar-expression
specifies the anchor point’s X coordinate.

Default 50

Interactions The DRAWSPACE= option determines the default interpretation of

the units for this setting. You can override the default with the
XSPACE= option.

If XSPACE=DATAVALUE, then this option's value is interpreted

using the XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted using
the primary X axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

Y=constant | scalar-expression
specifies the anchor point’s Y coordinate.

Default 50

Interactions The DRAWSPACE= option determines the default interpretation of

the units for this setting. You can override the default with the
YSPACE= option.

If YSPACE=DATAVALUE, then this option's value is interpreted

using the YAXIS= option.

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted using
the primary Y axis scale or to the secondary Y (Y2) axis scale.

Default Y
1226 Chapter 14 • Draw Statements

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

Details
A DRAWIMAGE statement draws an image in a graph. By default, the image is drawn
in the center of the graph. You can change the default position with the options
ANCHOR=, X=, Y=, XSPACE=, and YSPACE=. By default, the image is drawn in the
actual image size. You can change the default size with the WIDTH= and HEIGHT=
options.
For general information about the types of elements that can be drawn with the draw
statements, the drawing space and drawing units that they use, and how the drawn
elements are anchored, see Chapter 13, “Key Concepts for Using Draw Statements,” on
page 1189 . For detailed usage information, consult the SAS Graph Template Language:
User's Guide.

Example: DRAWIMAGE Statement

Example Graph
The following graph was generated by the “Example Program” on page 1227. The
example shows how to display an image in the bottom right corner of the graph wall.
DRAWLINE Statement 1227

Example Program
proc template;
define statgraph image;
begingraph;
entrytitle "Regression Fit Plot";
layout overlay;
modelband "myclm";
scatterplot x=height y=weight / primary=true;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
drawimage "c:\temp\saslogo.gif" /
anchor=bottomright x=98 y=2
drawspace=wallpercent ;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=image;

run;

DRAWLINE Statement
Draws a line from one point to another point.

Syntax
DRAWLINE X1=constant | scalar-expression
Y1=constant | scalar-expression
X2=constant | scalar-expression
Y2=constant | scalar-expression </option(s)>;

Summary of Optional Arguments

Appearance options
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line attributes for this draw statement.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted
using the primary X axis scale or to the secondary X (X2) axis scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted
using the primary Y axis scale or to the secondary Y (Y2) axis scale.
1228 Chapter 14 • Draw Statements

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.
X1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X1 value.
X2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X2 value
Y1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y1 value.
Y2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y2 value

ODS options
URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Required Arguments
X1=constant | scalar-expression
specifies the X value of the starting point for the line.

Interactions This value that is set for this option is interpreted using the
X1SPACE= option.

When X1SPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When X1SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y1=constant | scalar-expression
specifies the Y value of the starting point for the line.

Interactions This value that is set for this option is interpreted using the
Y1SPACE= option.

When Y1SPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When Y1SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.
DRAWLINE Statement 1229

X2=constant | scalar-expression
specifies the X value of the endpoint for the line.

Interactions This value that is set for this option is interpreted using the
X2SPACE= option.

When X2SPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When X2SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y2=constant | scalar-expression
specifies the Y value of the endpoint for the line.

Interactions This value that is set for this option is interpreted using the
Y2SPACE= option.

When Y2SPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When Y2SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Optional Arguments
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y values, or
both.

Default 0 (no offset, output is centered on the discrete X values, or the discrete
Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If the
Y axis is discrete, then a positive offset is up. If REVERSE=TRUE on
the X or Y axis, then the offset direction is also reversed.

Restriction This option applies only when the options X1SPACE=, X2SPACE=,
Y1SPACE=, or Y2SPACE= use DATAVALUE, and when X1, X2, Y1,
or Y2 are values on a discrete axis. For nondiscrete axes, this option is
ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT
1230 Chapter 14 • Draw Statements

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting for
only this draw statement.

This option sets the default drawing space, but individual settings in
the X or Y dimension can be overridden by the options X1SPACE=,
Y1SPACE=, X2SPACE=, or Y2SPACE=.

See “About the Drawing Space and Drawing Units” on page 1192

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT draws the output on top of the graph.

BACK draws the output behind the background areas, such as a layout or
legend background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer, you can
suppress the display of filled areas in the graph. You can also use
transparency to manage the element visibility. For more information, see
“About Drawing Layers” on page 1195 .

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes for this draw statement.

Default The GraphAnnoLine style element.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw statement is
selected.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
string | string-expression
specifies a valid HTML page reference (HREF) for the graphical element that is
drawn by this draw statement.

Example http://www.sas.com/technologies/analytics/
index.html
DRAWLINE Statement 1231

Requirement To generate selectable graphical elements, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted using
the primary X axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if X1SPACE=DATAVALUE.

X1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X1 value.

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the X1 value.

X2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X2 value

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the X2 value.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted using
the primary Y axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if Y1SPACE=DATAVALUE.

Y1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y1 value.

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the Y1 value.

Y2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y2 value

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the Y2 value.
1232 Chapter 14 • Draw Statements

Details
A DRAWLINE statement draws a line from a starting point that is specified with the X1
and Y1 arguments, to an ending point that is specified with the X2 and Y2 arguments.
DRAWLINE is similar to a DRAWARROW statement, using many of the same options,
but without the options for controlling the arrowhead(s).
For general information about the types of elements that can be drawn with the draw
statements, the drawing space and drawing units that they use, and how the drawn
elements are anchored, see Chapter 13, “Key Concepts for Using Draw Statements,” on
page 1189 . For detailed usage information, consult the SAS Graph Template Language:
User's Guide.

Example: DRAWLINE Statement

The following graph was generated by the “Example Program” on page 1232. The
example shows how to draw a diagonal reference line. One endpoint is point 0,0 and the
other is point 100,100. Both points are specified in the WALL area with PERCENT
units, making it easy to position the line without regard to the axis data ranges or the axis
offsets. To draw the line behind the series line and grid lines, you can set LAYER=
BACK and use the parent layout statement to turn off the display of the wall.

Example Program
proc template;
define statgraph diagonal;
begingraph;
entrytitle "Open vs. Close Price for Intel Stock 2003";
layout overlay / walldisplay=(outline)
xaxisopts=(griddisplay=on)
DRAWOVAL Statement 1233

yaxisopts=(griddisplay=on);
seriesplot x=open y=close / datalabel=date;
drawline x1=0 y1=0 x2=100 y2=100 /
x1space=wallpercent y1space=wallpercent
x2space=wallpercent y2space=wallpercent
lineattrs=GraphReference layer=back ;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.stocks template=diagonal;

where stock="Intel" and year(date)=2003;
format date monname3.;
run;

DRAWOVAL Statement
Draws an oval in the graph.

Syntax
DRAWOVAL X=constant | scalar-expression
Y=constant | scalar-expression
WIDTH=constant | scalar-expression
HEIGHT=constant | scalar-expression </option(s)>;

Summary of Optional Arguments

Appearance options
DISPLAY=STANDARD | ALL | (display-options)
specifies the features to display for the oval.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the fill attributes for this draw statement.
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the outline attributes for this draw statement.
ROTATE=number
specifies the angle of rotation for the oval, measured in degrees.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted
using the primary X axis scale or to the secondary X (X2) axis scale.
YAXIS=Y | Y2
1234 Chapter 14 • Draw Statements

specifies whether the data value for the arguments Y1 and Y2 are interpreted
using the primary Y axis scale or to the secondary Y (Y2) axis scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the Y= option.

Location options
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the oval.

ODS options
URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Size options
HEIGHTUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the HEIGHT=
option is interpreted as a percentage value, a pixel value, or a value that is in
the units of the data.
WIDTHUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the WIDTH= option
is interpreted as a percentage value, a pixel value, or a value that is in the
units of the data.

Required Arguments
X=constant | scalar-expression
specifies the X value of the anchor point.

Interactions The value that is set for this argument is interpreted using the
XSPACE= option.

When XSPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.
DRAWOVAL Statement 1235

Y=constant | scalar-expression
specifies the Y value of the anchor point.

Interactions The value that is set for this argument is interpreted using the
YSPACE= option.

When YSPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

WIDTH=constant | scalar-expression
specifies the width of the oval.

Interactions The value that is set for this argument is interpreted using the
WIDTHUNIT= and XSPACE= options.

When WIDTHUNIT=DATA, the value is interpreted using the

XAXIS= option.

HEIGHT=constant | scalar-expression
specifies the height of the oval.

Interactions The value that is set for this argument is interpreted using the
HEIGHTUNIT= and YSPACE= options.

When HEIGHTUNIT=DATA, the value is interpreted using the

YAXIS= option.

Optional Arguments
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the oval. The anchor point is relative to the unrotated
oval. The anchor point can be at the center of the oval or at eight points on the
bounding box of the rectangle. The following figure shows the anchor points for
TOPLEFT and CENTER.

The coordinates of the anchor point are set by the statement’s X and Y values, and by
the settings for the XSPACE= and YSPACE= options. The XAXIS= and YAXIS=
option might affect positioning when the XSPACE= or YSPACE= options are set to
DATAPIXEL, DATAPERCENT, or DATAVALUE.
When the oval is rotated, the anchor point remains relative to the unrotated oval
while the oval is rotated on its anchor point. See ROTATE= on page 1238.

Default CENTER
1236 Chapter 14 • Draw Statements

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y values, or
both.

Default 0 (no offset, output is centered on the discrete X values, or the discrete
Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If the
Y axis is discrete, then a positive offset is up. If REVERSE=TRUE on
the X or Y axis, then the offset direction is also reversed.

Restriction This option applies only when the options XSPACE= or YSPACE= use
DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DISPLAY=STANDARD | ALL | (display-options)

specifies the features to display for the oval.
STANDARD
displays an outlined oval.
ALL
displays an outlined, filled oval.
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays an outlined oval

FILL displays a filled oval

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the oval.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting for
only this draw statement.

This option sets the default drawing space, but individual settings in
the X or Y dimension can be overridden by the options XSPACE=,
YSPACE=, HEIGHTUNIT=, or WIDTHUNIT=.
DRAWOVAL Statement 1237

See “About the Drawing Space and Drawing Units” on page 1192

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the fill attributes for this draw statement.

Default The GraphAnnoShape style element

Tip The TRANSPARENCY= option sets the transparency for the fill and the
outline. You can combine this option with TRANSPARENCY= to set one
transparency for the outline but a different transparency for the fill.
Example:
transparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

HEIGHTUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the HEIGHT= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of the
data.

Default PERCENT

Interaction This setting combines with the YSPACE= setting to interpret the height
that is set in the HEIGHT= option.

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT draws the output on top of the graph.

BACK draws the output behind the background areas, such as a layout or
legend background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer, you can
suppress the display of filled areas in the graph. You can also use
transparency to manage the element visibility. For more information, see
“About Drawing Layers” on page 1195 .

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the outline attributes for this draw statement.

Default The GraphAnnoShape style element.

Interaction For this option to have any effect, the outline must be enabled by the
DISPLAY= option or by the ODS style that is in effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

1238 Chapter 14 • Draw Statements

ROTATE=number
specifies the angle of rotation for the oval, measured in degrees. The angle is
measured as if a horizontal line extended to the right through the oval anchor point as
shown in the following figure.

Positive angles rotate the oval counter clockwise, and negative angles rotate the oval
clockwise. The angle specification can exceed 360 degrees in absolute value.

Default 0. No rotation is performed

See ANCHOR= on page 1235

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the filled
area. You can combine this option with FILLATTRS= to set one
transparency for the outline but a different transparency for the fill. Here is
an example:
transparency=0.2 fillattrs=(transparency=0.6)

URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw statement is
selected.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
string | string-expression
specifies a valid HTML page reference (HREF) for the graphical element that is
drawn by this draw statement.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable graphical elements, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.
DRAWOVAL Statement 1239

WIDTHUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the WIDTH= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of the
data.

Default PERCENT

Interaction This setting combines with the XSPACE= setting to interpret the width
that is set in the WIDTH= option.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted using
the primary X axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted using
the primary Y axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

Details
A DRAWOVAL statement draws an oval in a graph. The oval position is determined by
the X and Y anchor points, and the size is determined by the HEIGHT and WIDTH
settings. You can manage the oval position with the options ANCHOR= , XSPACE= ,
and YSPACE= . You can manage the oval size with the HEIGHTUNIT= and
WIDTHUNIT= options.
For general information about the types of elements that can be drawn with the draw
statements, the drawing space and drawing units that they use, and how the drawn
elements are anchored, see Chapter 13, “Key Concepts for Using Draw Statements,” on
page 1189 . For detailed usage information, consult the SAS Graph Template Language:
User's Guide.
1240 Chapter 14 • Draw Statements

Example: DRAWOVAL Statement

The following graph was generated by the “Example Program” on page 1240. The
example uses DRAWOVAL to highlight a student’s data point. It draws an oval around
the marker symbol that represents the student’s height and weight, and it displays the
student’s name inside the oval. In the BEGINGRAPH statement, the setting for the
DRAWSPACE= option sets the drawing space and drawing units for the DRAWOVAL
and DRAWTEXT statements. In the DRAWOVAL statement, setting
TRANSPARENCY=0.75 ensures that the marker for Alfred is visible behind the oval.
The DRAWTEXT statement draws the text that identifies the student's name, using the
ANCHOR=, X=, and Y= options to position the text within the oval.

Example Program
proc template;
define statgraph drawoval;
begingraph / drawspace=datavalue;
entrytitle "Regression Fit Plot";
layout overlay;
modelband "myclm";
scatterplot x=height y=weight;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
drawoval x=69 y=112.5 width=15 height=20 /
display=all fillattrs=(color=green)
transparency=0.75 ;
drawtext "Alfred" / x=69 y=112 anchor=top;
endlayout;
endgraph;
DRAWRECTANGLE Statement 1241

end;

proc sgrender data=sashelp.class template=drawoval;

run;

DRAWRECTANGLE Statement
Draws a rectangle in the graph.

Syntax
DRAWRECTANGLE X=constant | scalar-expression
Y=constant | scalar-expression
WIDTH=constant | scalar-expression
HEIGHT=constant | scalar-expression </option(s)>;

Summary of Optional Arguments

Appearance options
CORNERRADIUS=number
specifies the radius of the rectangle corners.
DISPLAY=STANDARD | ALL | (display-options)
specifies the features to display for the rectangle.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the fill attributes for this draw statement.
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the outline attributes for this draw statement.
ROTATE=number
specifies the angle of rotation for the rectangle, measured in degrees.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted
using the primary X axis scale or to the secondary X (X2) axis scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted
using the primary Y axis scale or to the secondary Y (Y2) axis scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.
1242 Chapter 14 • Draw Statements

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the Y= option.

Location options
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the rectangle.

ODS options
URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Size options
HEIGHTUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the HEIGHT=
option is interpreted as a percentage value, a pixel value, or a value that is in
the units of the data.
WIDTHUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the WIDTH= option
is interpreted as a percentage value, a pixel value, or a value that is in the
units of the data.

Required Arguments
X=constant | scalar-expression
specifies the X value of the anchor point.

Interactions The value that is set for this argument is interpreted using the
XSPACE= option.

When XSPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y=constant | scalar-expression
specifies the Y value of the anchor point.

Interactions The value that is set for this argument is interpreted using the
YSPACE= option.

When YSPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.
DRAWRECTANGLE Statement 1243

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

WIDTH=constant | scalar-expression
specifies the width of the rectangle.

Interactions The value that is set for this argument is interpreted using the
WIDTHUNIT= and XSPACE= options.

When WIDTHUNIT=DATA, the value is interpreted using the

XAXIS= option.

HEIGHT=constant | scalar-expression
specifies the height of the rectangle.

Interactions The value that is set for this argument is interpreted using the
HEIGHTUNIT= and YSPACE= options.

When HEIGHTUNIT=DATA, the value is interpreted using the

YAXIS= option.

Optional Arguments
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the rectangle. The anchor point is relative to the
unrotated rectangle. The anchor point can be at the center of the rectangle or at eight
points on the rectangle border. The following figure shows the anchor points for
TOPLEFT and CENTER.

The coordinates of the anchor point are set by the statement’s X and Y values, and by
the settings for the XSPACE= and YSPACE= options. The XAXIS= and YAXIS=
option might affect positioning when the XSPACE= or YSPACE= options are set to
DATAPIXEL, DATAPERCENT, or DATAVALUE.
When the rectangle is rotated, the anchor point remains relative to the unrotated
rectangle while the rectangle is rotated on its anchor point. See ROTATE= on page
1246.

Default CENTER

CORNERRADIUS=number
specifies the radius of the rectangle corners.

Default 0

Range 0–1, where 0 specifies square corners and 1 specifies the most rounded
corners
1244 Chapter 14 • Draw Statements

Example cornerradius=0.2

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y values, or
both.

Default 0 (no offset, output is centered on the discrete X values, or the discrete
Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If the
Y axis is discrete, then a positive offset is up. If REVERSE=TRUE on
the X or Y axis, then the offset direction is also reversed.

Restriction This option applies only when the options XSPACE= or YSPACE= use
DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DISPLAY=STANDARD | ALL | (display-options)

specifies the features to display for the rectangle.
STANDARD
displays an outlined rectangle.
ALL
displays an outlined, filled rectangle.
(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

OUTLINE displays an outlined rectangle

FILL displays a filled rectangle

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the rectangle.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting for
only this draw statement.
DRAWRECTANGLE Statement 1245

This option sets the default drawing space, but individual settings in
the X or Y dimension can be overridden by the options XSPACE=,
YSPACE=, HEIGHTUNIT=, or WIDTHUNIT=.

See “About the Drawing Space and Drawing Units” on page 1192

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the fill attributes for this draw statement.

Default The GraphAnnoShape style element

Tip The TRANSPARENCY= option sets the transparency for the fill and the
outline. You can combine this option with TRANSPARENCY= to set one
transparency for the outline but a different transparency for the fill.
Example:
transparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

HEIGHTUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the HEIGHT= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of the
data.

Default PERCENT

Interaction This setting combines with the YSPACE= setting to interpret the height
that is set in the HEIGHT= option.

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT draws the output on top of the graph.

BACK draws the output behind the background areas, such as a layout or
legend background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer, you can
suppress the display of filled areas in the graph. You can also use
transparency to manage the element visibility. For more information, see
“About Drawing Layers” on page 1195 .

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the outline attributes for this draw statement.

Default The GraphAnnoShape style element.

Interaction For this option to have any effect, the outline must be enabled by the
DISPLAY= option or by the ODS style that is in effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.
1246 Chapter 14 • Draw Statements

“Line Options” on page 1349 for available line-options.

ROTATE=number
specifies the angle of rotation for the rectangle, measured in degrees. The angle is
measured as if a horizontal line extended to the right through the rectangle anchor
point as shown in the following figure.

Positive angles rotate the rectangle counter clockwise, and negative angles rotate the
rectangle clockwise. The angle specification can exceed 360 degrees in absolute
value.

Default 0. No rotation is performed

See ANCHOR= on page 1243

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the filled
area. You can combine this option with FILLATTRS= to set one
transparency for the outline but a different transparency for the fill. Here is
an example:
transparency=0.2 fillattrs=(transparency=0.6)

URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw statement is
selected.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
string | string-expression
specifies a valid HTML page reference (HREF) for the graphical element that is
drawn by this draw statement.

Example http://www.sas.com/technologies/analytics/
index.html
DRAWRECTANGLE Statement 1247

Requirement To generate selectable graphical elements, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

WIDTHUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the WIDTH= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of the
data.

Default PERCENT

Interaction This setting combines with the XSPACE= setting to interpret the width
that is set in the WIDTH= option.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted using
the primary X axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted using
the primary Y axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

Details
A DRAWRECTANGLE statement draws a rectangle in a graph. The rectangle position
is determined by the X and Y anchor points, and the size is determined by the HEIGHT
and WIDTH settings. You can manage the rectangle position with the options
ANCHOR= , XSPACE= , and YSPACE= . You can manage the rectangle size with the
HEIGHTUNIT= and WIDTHUNIT= options.
For general information about the types of elements that can be drawn with the draw
statements, the drawing space and drawing units that they use, and how the drawn
elements are anchored, see Chapter 13, “Key Concepts for Using Draw Statements,” on
1248 Chapter 14 • Draw Statements

page 1189 . For detailed usage information, consult the SAS Graph Template Language:
User's Guide.

Example: DRAWRECTANGLE Statement

The following graph was generated by the “Example Program” on page 1248. The
example uses DRAWRECTANGLE to highlight a student’s data point. It draws a
rectangle around the marker symbol that represents the student’s height and weight, and
it displays the student’s name inside the rectangle. In the BEGINGRAPH statement, the
setting for the DRAWSPACE= option sets the drawing space and drawing units for the
DRAWRECTANGLE and DRAWTEXT statements. In the DRAWRECTANGLE
statement, setting TRANSPARENCY=0.75 ensures that the marker for Alfred is visible
behind the rectangle. The DRAWTEXT statement draws the text that identifies the
student's name, using the ANCHOR=, X=, and Y= options to position the text within the
rectangle.

Example Program
proc template;
define statgraph drawrectangle;
begingraph / drawspace=datavalue;
entrytitle "Regression Fit Plot";
layout overlay;
modelband "myclm";
scatterplot x=height y=weight;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
drawrectangle x=69 y=112.5 width=10 height=15 /
display=all fillattrs=(color=green)
transparency=0.75 ;
DRAWTEXT Statement 1249

drawtext "Alfred" / x=69 y=112 anchor=top;

endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=drawrectangle;

run;

DRAWTEXT Statement
Draws and anchors in a graph a text box that contains one or more lines of formatted text.

Syntax
DRAWTEXT text-item <text-item …> </option(s)>;

Summary of Optional Arguments

Appearance options
BORDER=TRUE | FALSE
specifies whether a border is drawn around the text box.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the border line attributes for this draw statement.
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the text box’s
border.
ROTATE=number
specifies the angle of rotation of the text box, measured in degrees.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted
using the primary X axis scale or to the secondary X (X2) axis scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted
using the primary Y axis scale or to the secondary Y (Y2) axis scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.
1250 Chapter 14 • Draw Statements

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that
is specified in the Y= option.

Location options
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the text box.
JUSTIFY=LEFT | CENTER | RIGHT
specifies the alignment of text that wraps within the text box.
X=constant | scalar-expression
specifies the anchor point’s X coordinate.
Y=constant | scalar-expression
specifies the anchor point’s Y coordinate.

ODS options
URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Size options
WIDTH=positive-number
specifies the width of the text box.
WIDTHUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the WIDTH= option
is interpreted as a percentage value, a pixel value, or a value that is in the
units of the data.

Required Argument
text-item <…text-item>
specifies one or more pieces of text for the text box. Each text-item has the following
form:
<prefix-option> "string" | dynamic | character-expression | {text-command}
Each piece of text can have a prefix setting that precedes the piece of text. A piece of
text is either a string literal, a dynamic, or a text command. All text-items are
concatenated into one string, which might be wrapped, based on the settings for the
WIDTH= and WIDTHUNIT= settings. Leading and trailing blanks in the
concatenated string are always used.
When used, a prefix option applies to the immediately following piece of text and
also to all subsequent text strings and text-commands until another prefix option is
specified.

Requirements string must be enclosed in quotation marks.

DRAWTEXT Statement 1251

character-expression must be enclosed in an EVAL function.

text-command must be enclosed in braces.

Note Leading a trailing spaces are preserved.

See Chapter 10, “Managing Text Items,” on page 1137

Optional Arguments
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the text box. The anchor point is relative to the
unrotated text. It can be at the center of the text box or at eight points on the border
of the text box bounding box. The following figure shows the anchor points for
TOPLEFT and LEFT.

The coordinates of the anchor point are set by the X= and Y= options, and by the
XSPACE= and YSPACE= options. The XAXIS= and YAXIS= option might affect
positioning when the XSPACE= or YSPACE= options are set to DATAPIXEL,
DATAPERCENT, or DATAVALUE.
The text box has a fixed width, determined by the WIDTH= and WIDTHUNIT=
options. The height of the text box is based on the amount of text specified and the
font size. The height of the text grows in a direction that is related to the anchor
point. For example, if ANCHOR=TOPLEFT, then the text box height extends
downward from the anchor point and its width extends to the right. If
ANCHOR=CENTER, then half the text box width and half the text box height
extend equally left and right, as well as top to bottom from the anchor point. If
ANCHOR=BOTTOM, the text box height extends upward from the anchor point and
the text box width is centered at the anchor point.
When the text is rotated, the anchor point remains relative to the unrotated text box
while the text box is rotated on its anchor point. See ROTATE= on page 1253.

Default CENTER

BORDER=TRUE | FALSE
specifies whether a border is drawn around the text box.

Default FALSE

Tip Use the BORDERATTRS= option to control the appearance of the border.

See “boolean ” on page 1339 for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the border line attributes for this draw statement.

Default The GraphBorderLines style element.

1252 Chapter 14 • Draw Statements

Interaction BORDER=TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y values, or
both.

Default 0 (no offset, output is centered on the discrete X values, or the discrete
Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If the
Y axis is discrete, then a positive offset is up. If REVERSE=TRUE on
the X or Y axis, then the offset direction is also reversed.

Restriction This option applies only when the options XSPACE= or YSPACE= use
DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting for
only this draw statement.

This option sets the default drawing space, but individual settings in
the X or Y dimension can be overridden by the options XSPACE=,
YSPACE=, HEIGHTUNIT=, or WIDTHUNIT=.

See “About the Drawing Space and Drawing Units” on page 1192

JUSTIFY=LEFT | CENTER | RIGHT

specifies the alignment of text that wraps within the text box.
LEFT
forces the first character of each line to appear at the left margin (distance from
the left border plus the left pad amount).
CENTER
forces each line to be centered in the text box between the left and right pad
amounts.
DRAWTEXT Statement 1253

RIGHT
forces the last character of each line to appear at the right margin (distance from
the right border minus the right pad amount).

Default LEFT

Interaction Text is wrapped based on the width of the specified text, the font size,
and the setting in the PAD= option.

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT draws the output on top of the graph.

BACK draws the output behind the background areas, such as a layout or
legend background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer, you can
suppress the display of filled areas in the graph. You can also use
transparency to manage the element visibility. For more information, see
“About Drawing Layers” on page 1195 .

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the text box’s border.
dimension
specifies a dimension to use for the extra space at the left, right, top, and bottom
of the text box border.
(pad-options)
a space-separated list of one or more of the following name-value-pair options
enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the left

side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.

Note Sides that are not assigned padding are padded with the default amount.

Tip Use pad-options to create non-uniform padding.

Default (LEFT=3 RIGHT=3 TOP=0 BOTTOM=0)

Note The default units for dimension are pixels.

See “dimension” on page 1340

ROTATE=number
specifies the angle of rotation of the text box, measured in degrees. The text box is
rotated around its anchor point. The angle is measured from a horizontal line passing
1254 Chapter 14 • Draw Statements

through the anchor point of the text box to the right. The following figure shows the
rotation of text around a top anchor point.

Positive angles rotate the text box counter clockwise, and negative angles rotate the
text box clockwise. The angle specification can exceed 360 degrees in absolute
value.

Default 0. No rotation is performed

See ANCHOR= on page 1251

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

URL=string | string-expression
specifies an HTML page that is displayed when the output of this draw statement is
selected.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
string | string-expression
specifies a valid HTML page reference (HREF) for the graphical element that is
drawn by this draw statement.

Example http://www.sas.com/technologies/analytics/
index.html

Requirement To generate selectable graphical elements, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

WIDTH=positive-number
specifies the width of the text box.

Default 10

Interaction This option's value is interpreted using the WIDTHUNIT= option.

DRAWTEXT Statement 1255

WIDTHUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the WIDTH= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of the
data.

Default PERCENT

Interaction This setting combines with the XSPACE= setting to interpret the width
that is set in the WIDTH= option.

X=constant | scalar-expression
specifies the anchor point’s X coordinate.

Default 50

Interactions The DRAWSPACE= option determines the default interpretation of

the units for this setting. You can override the default with the
XSPACE= option.

If XSPACE=DATAVALUE, then this option's value is interpreted

using the XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 are interpreted using
the primary X axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

Y=constant | scalar-expression
specifies the anchor point’s Y coordinate.

Default 50

Interactions The DRAWSPACE= option determines the default interpretation of

the units for this setting. You can override the default with the
YSPACE= option.

If YSPACE=DATAVALUE, then this option's value is interpreted

using the YAXIS= option.

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.
1256 Chapter 14 • Draw Statements

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 are interpreted using
the primary Y axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT |
DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

Prefix Option
TEXTATTRS=style-element | style-element (text-options) | (text-options)
prefix-option that specifies the color and font properties of the entire text string or
individual text-items.

Default The GraphValueText style element.

Interaction When multiple TEXTATTRS= prefix options are used, each one
cancels the last, resetting all text properties to the default set by the
GraphValueText style element. Subsequent text-items to the right are
then assigned the text properties specified in the closest TEXTATTRS=
setting to their left. Thus, to vary the text properties across text-items,
you do not have to override settings from a previous TEXTATTRS=
setting. Each TEXTATTRS= specification resets all text properties to
the default so that only the new settings are applied to subsequent text-
items.

Tip To ensure that all text has the same text properties, use this prefix
option once only and place it before the first text-item.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Text Options” on page 1351 for available text-options.

Text Commands
{ SUB "string" | dynamic }
text-command that specifies that the string or dynamic is to appear as subscript text.

See “Rules for Unicode and Special Character Specifications” on page 1142

Example drawtext "y = " b{sub "0"} " + b" {sub "1"} "x";

{ SUP "string" | dynamic }

text-command that specifies that the string or dynamic is to appear as superscript
text.

See “Rules for Unicode and Special Character Specifications” on page 1142
Example: DRAWTEXT Statement 1257

Example drawtext "R" {sup "2"} " = " {format (6.4) RSQUARED} ;

{ UNICODE "hex-string"x | keyword | dynamic }

text-command that specifies a glyph (character) to be displayed using its Unicode
specification or keyword equivalent.
"hex-string"x
a four-byte hexadecimal constant that represents a UNICODE character in the
current font. For a complete listing, see http://unicode.org/charts/charindex.html.
keyword
a SAS keyword for a UNICODE character. For a listing of SAS supplied
keywords, see Appendix 2, “Reserved Keywords and Unicode Values,” on page
1343.
dynamic
a dynamic variable that resolves to either "hex-string"x or a keyword for a
UNICODE character.

Note The UNICODE text command attempts to access a UNICODE value in

the current font. Not all fonts support accessing characters through their
UNICODE value. Some fonts support only a limited set of UNICODE
values. If the UNICODE value is not accessible, then the command might
be ignored or a nonprintable character might be substituted.

See “Using UNICODE Text” on page 1141

“Rules for Unicode and Special Character Specifications” on page 1142

Example The following statements show how to use the {UNICODE} text
command:
drawtext {unicode alpha} "=" CONF;

drawtext {unicode "03B1"x} "=" CONF;

Details
A DRAWTEXT statement draws a text box that contains one or more lines of text. The
text can be formatted, using the TEXTATTRS= prefix option. By default, the text box is
drawn in the center of the graph. You can change the default position with the options
ANCHOR= , X= , Y= , XSPACE= , and YSPACE= .
For general information about the types of elements that can be drawn with the draw
statements, the drawing space and drawing units that they use, and how the drawn
elements are anchored, see Chapter 13, “Key Concepts for Using Draw Statements,” on
page 1189 . For detailed usage information, consult the SAS Graph Template Language:
User's Guide.

Example: DRAWTEXT Statement

The following graph was generated by the “Example Program” on page 1258. The first
DRAWTEXT statement shows how to draw multiple lines of text in a specific position
within the graph. The second DRAWTEXT statement shows how to create a watermark,
which is achieved by applying transparency to text that is rotated within the graph.
1258 Chapter 14 • Draw Statements

Example Program
proc template;
define statgraph modelfit;
begingraph;
entrytitle "Regression Fit Plot";
layout lattice;
layout overlay / xaxisopts=(offsetmin=0.1);
drawtext textattrs=(style=italic size=8pt)
"Band shows 99% Confidence Limit of Mean" /
anchor=bottomleft width=15 widthunit=percent
xspace=wallpercent yspace=wallpercent
x=0 y=10 justify=center ;
modelband "myclm";
scatterplot x=height y=weight / primary=true;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
endlayout;
endlayout;
drawtext textattrs=(color=gray size=52pt) "CONFIDENTIAL" /
transparency=0.75 rotate=-35
width=110 widthunit=percent justify=center ;
endgraph;
end;

proc sgrender data=sashelp.class template=modelfit;

run;
1259

Part 10

GTL Annotation Facility

Chapter 15
About the GTL Annotation Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1261

Chapter 16
The ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1267
1260
1261

Chapter 15
About the GTL Annotation Facility

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1261
The Annotation Data Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1262
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1262
Annotation Data Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1262
About the Coordinate Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1264
About the Drawing Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1264
Using the SGANNO Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1265
The ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1266
The SGRENDER Statement SGANNO Option . . . . . . . . . . . . . . . . . . . . . . . . . . . 1266

Overview
The GTL supports data-set-driven annotations, which enables you to add the following
graphics elements to your graphs:
• text
• lines and arrows
• circles and ovals
• squares and rectangles
• polygons and polylines
• images
Unlike graphics elements that are drawn using GTL draw statements in a GTL template,
data-set-driven annotations are drawn from graphics instructions that are stored in a SAS
data set. The GTL annotation facility is similar to the SAS/GRAPH Annotate facility.
The GTL annotation facility enables you to separate your annotation instructions from
your template statements. To change your annotations, you can specify a different
annotation data set or modify the instructions in the original data set. You do not have to
modify your template code.
The following items are required in order to use the GTL annotation facility:
• a SAS data set that contains the annotation instructions
• at least one ANNOTATE statement in your GTL template
1262 Chapter 15 • About the GTL Annotation Facility

• an SGRENDER statement that includes the SGANNO=data-set-name option, which

specifies the name of the data set that contains the annotation instructions.
This section describes the requirements for using the GTL annotation facility. For
examples of how to use the facility to annotate your graphs, see “Adding Data-Driven
Annotations to Your Graph” in SAS Graph Template Language: User's Guide.

The Annotation Data Set

Overview
The annotation data set contains the instructions for drawing annotations. Each
observation in the data set contains columns for an annotation instruction. The following
information is typically stored in these columns:
• the annotation function name
• the coordinates of the annotation location
• the drawing space
• the attributes of the annotation, such as color, font, and so on
The information that is required for each instruction is function-dependent. See
“Annotation Data Requirements” on page 1262.
To create the annotation data set, you can use the same methods that you would use to
create any SAS data set. The most common method is to use a DATA step. For more
information about the DATA step, see SAS Language Reference: Concepts. Macros are
available that you can run in your DATA step to create the observations for your
annotations. See “Using the SGANNO Macros” on page 1265.

Annotation Data Requirements

Each observation in the annotation data set must provide sufficient information to
complete an annotation instruction. To provide this information, certain columns must be
included in each observation. All instructions require a Function column, which is a
character column that stores the name of the function that is to be performed. An
annotation function draws an annotation, such as a line, arrow, oval, and so on. The
Function column can specify one of the following function names:

ARROW POLYGON TEXT

IMAGE POLYLINE TEXTCONT
LINE POLYCONT
OVAL RECTANGLE
For more information about these functions, see “SG Annotation Function Dictionary”
in SAS ODS Graphics: Procedures Guide.
Each function requires one or more additional columns that provide required information
such as the location coordinates, dimensions, and so on. The following table lists the
additional columns that are required for each function. At a minimum, you must include
the Function column and the columns listed in the following table in each observation in
your annotation data set.
The Annotation Data Set 1263

Additional Columns
FUNCTION Required* Description

ARROW X1 or XC1 These columns store the starting coordinates

(numeric or character).
Y1 or YC1

X2 or XC2 These columns store the ending coordinates

(numeric or character).
Y2 or YC2

IMAGE Image This column stores the path to the image file
(character).

LINE X1 or XC1 These columns store the starting coordinates

(numeric or character).
Y1 or YC1

X2 or XC2 These columns store the ending coordinates

(numeric or character).
Y2 or YC2

OVAL Height These columns store the oval dimensions

(numeric).
Width

X1 or XC1 These columns store the oval anchor

coordinates (numeric or character).
Y1 or YC1

POLYGON X1 or XC1 These columns store the polygon starting

coordinates (numeric or character).
Y1 or YC1

POLYLINE X1 or XC1 These columns store the polyline starting

coordinates (numeric or character).
Y1 or YC1

POLYCONT X1 or XC1 These columns store the ending coordinates of

a polygon or polyline segment (numeric or
Y1 or YC1
character).

RECTANGLE Height These columns store the rectangle dimensions

(numeric).
Width

X1 or XC1 These columns store the rectangle anchor

coordinates (numeric).
Y1 or YC1

TEXT and Label This column stores the annotation text

TEXTCONT (character).

* See “About the Coordinate Columns” on page 1264.

For the remaining information such as LINECOLOR, TEXTFONT, and so on, default
values are used. To change the default values, you can add the necessary columns to
your instruction observations. For information about other columns that you can add for
each function, see “SG Annotation Function Dictionary” in SAS ODS Graphics:
Procedures Guide
1264 Chapter 15 • About the GTL Annotation Facility

About the Coordinate Columns

The annotation coordinates specify the location of the annotation as X and Y values. In
some cases, such as for a line or arrow, two sets of coordinates are required in order to
specify the beginning and ending locations of the annotation. You can specify numeric or
character columns for each coordinate value. For numeric coordinate values, the location
or starting location coordinates are stored in the X1 and Y1 columns in the annotation
data set. When ending coordinates are required, the numeric coordinates are stored in the
X2 and Y2 columns. For character values, the coordinates are stored in the XC1 and
YC1 columns, and when required, the XC2 and YC2 columns.
Note: If both the numeric and character columns are specified for a coordinate value,
then the numeric column takes precedence and the character column is ignored.

About the Drawing Space

When you specify X and Y coordinates for an annotation, the coordinate values refer to
a drawing space. By default, this space is a percentage of the graph area. Specifying
X=50 and Y=50, for example, refers to a point in the center of the graph area. Here is a
complete list of the drawing spaces that you can use.

DATAPERCENT GRAPHPERCENT LAYOUTPIXEL

DATAPIXEL GRAPHPIXEL WALLPERCENT
DATAVALUE LAYOUTPERCENT WALLPIXEL
For more information about these drawing spaces, see “About the Drawing Space and
Drawing Units” on page 1192.
To change the default drawing space for your annotations, you can add the columns
shown in the following table to your annotation observations.

Column Name Description

DrawSpace Specifies the drawing space for all coordinates. This value cannot be used
with the POLYCONT function.

X1Space Specifies the drawing space for the X1 coordinate. This value cannot be
used with the TEXTCONT function.

X2Space Specifies the drawing space for the ending X2 coordinate. This value can be
used with the ARROW and LINE functions only.

Y1Space Specifies the drawing space for the Y1 coordinate. This value cannot be
used with the TEXTCONT function.

Y2Space Specifies the drawing space for the ending Y2 coordinate. This value can be
used with the ARROW and LINE functions only.

For information about these columns, see “SG Annotation Function Dictionary” in SAS
ODS Graphics: Procedures Guide
The Annotation Data Set 1265

Using the SGANNO Macros

Starting with the first maintenance release of SAS 9.4, macros are available that you can
run in a DATA step to simplify the process of creating your SG annotation data sets. The
following table lists the available macros.

Table 15.1 SG Annotation Macros

Macro Name Description

%SGANNO Compiles the available macros and makes them available for
you to use.
Note: You must run the %SGANNO macro in your SAS
session before you can use any of the other annotation macros
that are listed in this table.

%SGANNO_HELP Displays help information for a specified annotation macro.

%SGARROW Creates an observation that draws an arrow.

%SGIMAGE Creates an observation that displays an image.

%SGLINE Creates an observation that draws a line.

%SGOVAL Creates an observation that draws an oval.

%SGPOLYCONT Creates an observation that draws a polygon or polyline

segment.

%SGPOLYGON Creates an observation that specifies the starting point of a

polygon.
Note: Use the %SGPOLYCONT macro to draw the segments.

%SGPOLYLINE Creates an observation that specifies the starting point of a

polyline.
Note: Use the %SGPOLYCONT macro to draw the segments.

%SGRECTANGLE Creates an observation that draws a rectangle.

%SGTEXT Creates an observation that draws a single line of text or the

first line of text in a multiline annotation.
Note: For multiple lines of text, use the %SGTEXTCONT
macro for the subsequent lines.

%SGTEXTCONT Creates an observation that draws a continuing line of text in a

multiline annotation.

For more information about these macros, see SAS ODS Graphics: Procedures Guide.
1266 Chapter 15 • About the GTL Annotation Facility

The ANNOTATE Statement

To render the annotations in your annotation data set, you must include at least one
ANNOTATE statement in the GTL template for your graph. You can place the
ANNOTATE statement anywhere in the template to render your annotations. The
annotations are drawn using the context in which the ANNOTATE statement is
encountered. You can use the ANNOTATE statement with no options to render all of the
annotations in your annotation data set. You can use the ID= option in the ANNOTATE
statement to render only a subset of the annotations.
For more information about the ANNOTATE statement, see “ANNOTATE Statement”
on page 1267.

The SGRENDER Statement SGANNO Option

The SGRENDER statement SGANNO= option specifies the name of a SAS data set that
contains annotation instructions. To render your graph with annotations, you must use
the SGANNO= option in the SGRENDER statement to specify the name of your
annotation data set. When the ANNOTATE statement is encountered in the template, the
annotation instructions are read from the annotation data set and are rendered in the
current context in the graph. If the template does not include the ANNOTATE statement,
the annotations in the annotation data set are ignored. See “SGANNO=annotation-data-
set” in SAS ODS Graphics: Procedures Guide.
1267

Chapter 16
The ANNOTATE Statement

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1267
ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1267

Dictionary

ANNOTATE Statement
Draws annotations from annotation instructions that are stored in a SAS data set.
Requirements: The annotation instructions must be stored in a SAS data set.
The annotation data set must be specified by the SGANNO= option in the
SGRENDER statement.
Tips: When the ANNOTATE statement is placed inside a LAYOUT container, the layout
container is used as the context when DRAWSPACE for an annotation is specified
as LAYOUT, WALL, or DATA.
When the DRAWSPACE for an annotation object is GRAPH, the context for drawing
the object is the graph area regardless of where the ANNOTATE statement is
located.
The LAYER option determines whether the annotations are drawn in front of the
graph elements or behind them.
See: “The Annotation Data Set” on page 1262
“SG Annotation Function Dictionary” in SAS ODS Graphics: Procedures Guide
SGRENDER statement SGANNO= option
“Adding Data-Driven Annotations to Your Graph” in SAS Graph Template Language:
User's Guide

Syntax
ANNOTATE </ID="annotation-ID">
1268 Chapter 16 • The ANNOTATE Statement

Optional Argument
ID="annotation-identifier"
specifies the ID column value of the annotations that are to be drawn. The ID column
of the annotation data set contains a unique character value that identifies the subset
to which each annotation belongs. All annotations in the annotation data set with an
ID column value that matches the specified annotation identifier are drawn. If the
annotation data set does not contain an ID column or if no ID column value matches
the specified identifier, then no annotations are drawn.

Default All of the annotations in the annotation data set are drawn.

Requirement The annotation data set must contain an ID column, and at least one
ID column value must match the specified identifier value.

Tips The ID= option can be used with the annotation data set ID column to
subset the annotations that are stored in an annotation data set.

The identifier value matching is case sensitive.

See “Subsetting Annotations” in SAS Graph Template Language: User's

Guide

Details
You can use the ANNOTATE statement anywhere in a GTL template to render
annotation objects in the current context. The annotation objects are read from an
annotation data set that is specified by the SGANNO= option in the SGRENDER
statement. You must use at least one ANNOTATE statement in the template to draw the
annotations. By default, all of the annotations in the annotation data set are drawn. You
can use the ID column in the annotation data set with the ID= option in the ANNOTATE
statement to draw only a subset of the annotations in the data set, if you want to. See
“Subsetting Annotations” in SAS Graph Template Language: User's Guide.

Examples

Example 1: Adding Simple Text Annotations

Here is an example that uses the GTL annotation facility to place custom Y-axis labels
inside the plot wall for each of two side-by-side plots. Each label is rotated 90 degrees
and is placed on the inside edge of the plot wall near the plot’s Y axis. The default Y-axis
labels are suppressed.
Example 1: Adding Simple Text Annotations 1269

Here is the output for this example.

Example Program
/* Create the annotation data set */
data anno;
length label $30 id $5 anchor $12;
drawspace="wallpercent";

/* Create the bar chart Y-axis label. */

id='BAR'; function='text'; x1=0;
textweight='bold'; anchor='top'; rotate=90;
width=1000; widthunit="pixel";
label="MPG (City)";
output;

/* Create the histogram Y-axis label. */

id='HIST'; anchor='bottom'; x1=100;
label="Distribution of MPG (Percent)";
output;
run;

/* Define the template */

proc template;
define statgraph anno;
begingraph;
entrytitle "Vehicle Statistics";
layout lattice / columns=2 columngutter=10;
layout overlay /
xaxisopts=(offsetmin=0.2 offsetmax=0.2)
yaxisopts=(display=(ticks tickvalues));

/* Draw the barchart of origin and MPG city */

barchart x=origin y=mpg_city / name="bar" stat=mean
group=type groupdisplay=cluster clusterwidth=0.7
dataskin=sheen;
discretelegend "bar";
1270 Chapter 16 • The ANNOTATE Statement

annotate / id="BAR"; /* Draw the barchart label. */

endlayout;
layout overlay /
y2axisopts=(display=(ticks tickvalues));

/* Draw the histogram of MPG city */

histogram mpg_city / dataskin=sheen yaxis=y2;
densityplot mpg_city / yaxis=y2;
annotate / id="HIST"; /* Draw the histogram label. */
endlayout;
endlayout;
endgraph;
end;
run;

/* Render the graph with the annotation */

proc sgrender data=sashelp.cars template=anno sganno=anno;
where type in ("Sedan" "Sports" "SUV");
run;

Program Description

Create the annotation data set Anno. The annotation data set contains two
observations, one for each label. The DRAWSPACE is set to WALLPERCENT for both
labels. The ID column specifies an identifier for each label. The bar chart label is
identified as BAR. It is rotated 90 degrees, anchored on TOP, and is placed at 0% of the
wall space along the X axis. The default value 50% percent is used along the Y axis to
center the label vertically. This places the label on inside left edge of the bar chart wall.
This histogram label is identified as HIST. It is rotated 90 degrees, anchored on
BOTTOM, and is placed at 100% of the wall space along the X axis. Like the bar chart
label, the default 50% percent is used along the Y axis. This places the label on the right
inside edge of the histogram plot wall.
/* Create the annotation data set */
data anno;
length label $30 id $5 anchor $12;
drawspace="wallpercent";

/* Create the bar chart Y-axis label. */

id='BAR'; function='text'; x1=0;
textweight='bold'; anchor='top'; rotate=90;
width=1000; widthunit="pixel";
label="MPG (City)";
output;

/* Create the histogram Y-axis label. */

id='HIST'; anchor='bottom'; x1=100;
label="Distribution of MPG (Percent)";
output;
run;

Define the graph template. The template defines a two-column, one-row lattice for the
two plots. For both plots, the DISPLAY= Y axis option suppresses the default axis label.
For the bar chart, the OFFSETMIN= and OFFSETMAX= X axis options reserve space
for the Y-axis label. The ANNOTATE statement draws the axis label. The ID="BAR"
option in the ANNOTATE statement draws the bar chart Y-axis label in the context of
Example 2: Using the SG Annotation Macros to Generate Your Annotation Data 1271

the bar chart’s overlay layout. For the histogram, the ANNOTATE statement draws the
histogram Y-axis label (ID="HIST") in the context of the histogram’s overlay layout.
/* Define the template */
proc template;
define statgraph anno;
begingraph;
entrytitle "Vehicle Statistics";
layout lattice / columns=2 columngutter=10;
layout overlay /
xaxisopts=(offsetmin=0.2 offsetmax=0.2)
yaxisopts=(display=(ticks tickvalues));

/* Draw the barchart of origin and MPG city */

barchart x=origin y=mpg_city / name="bar" stat=mean
group=type groupdisplay=cluster clusterwidth=0.7
dataskin=sheen;
discretelegend "bar";
annotate / id="BAR"; /* Draw the barchart label. */
endlayout;
layout overlay /
y2axisopts=(display=(ticks tickvalues));

/* Draw the histogram of MPG city */

histogram mpg_city / dataskin=sheen yaxis=y2;
densityplot mpg_city / yaxis=y2;
annotate / id="HIST"; /* Draw the histogram label. */
endlayout;
endlayout;
endgraph;
end;
run;

Render the graph with the annotations. The SGANNO=ANNO option is included in
the SGRENDER statement to specify the name of the annotation data set. When the
graph is rendered, the ANNOTATE statements in the graph template cause the
annotations to be drawn.
/* Render the graph with the annotation */
proc sgrender data=sashelp.cars template=anno sganno=anno;
where type in ("Sedan" "Sports" "SUV");
run;

Example 2: Using the SG Annotation Macros to Generate Your

Annotation Data
In “Example 1: Adding Simple Text Annotations” on page 1268, standard SAS DATA
step statements are used to create the observations for the text annotations. This example
shows you how to use the SG annotation macros instead to create the same data set.

Example Program
/* Create the annotation data set */
data anno;
/* Compile the annotation macros */
%sganno;

/* Create the bar chart Y-axis label. */

1272 Chapter 16 • The ANNOTATE Statement

%sgtext(
reset=all,
id=BAR,
drawspace=wallpercent,
x1=0,
label="MPG (City)",
textweight=bold,
anchor=top,
rotate=90,
width=1000,
widthunit=pixel);

/* Create the histogram Y-axis label. */

%sgtext(
id=HIST,
x1=100,
label="Distribution of MPG (Percent)",
anchor=bottom);
run;

Program Description
Here is the SAS code.

Open the DATA step and run the %SGANNO macro. After you open the DATA step,
you must run the %SGANNO macro to compile the SG annotation macros and make
them available for your use. You need to compile the macros only once during a SAS
session.
/* Create the annotation data set */
data anno;
/* Compile the annotation macros */
%sganno;

Create the observation for the bar chart Y-axis label. Use the %SGTEXT macro to
generate the observation for the bar chart Y-axis label.
/* Create the bar chart Y-axis label. */
%sgtext(
reset=all,
id=BAR,
drawspace=wallpercent,
x1=0,
label="MPG (City)",
textweight=bold,
anchor=top,
rotate=90,
width=1000,
widthunit=pixel);

Create the histogram Y-axis label, and close the DATA step. Specify the new values
for the ID, X1, Label, and Anchor columns. To carry over the remaining column values
from the previous macro call, do not specify RESET=ALL.
/* Create the histogram Y-axis label. */
%sgtext(
id=HIST,
x1=100,
Example 2: Using the SG Annotation Macros to Generate Your Annotation Data 1273

label="Distribution of MPG (Percent)",

anchor=bottom);
run;

See Also
• Chapter 15, “About the GTL Annotation Facility,” on page 1261
• “Adding Data-Driven Annotations to Your Graph” in SAS Graph Template
Language: User's Guide
1274 Chapter 16 • The ANNOTATE Statement
1275

Part 11

Attribute Maps

Chapter 17
Key Concepts for Using Attribute Maps . . . . . . . . . . . . . . . . . . . . . . . . 1277

Chapter 18
Discrete Attribute Map Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1287

Chapter 19
Range Attribute Map Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1301
1276
1277

Chapter 17
Key Concepts for Using Attribute
Maps

About Attribute Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1277

Defining a Discrete Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1278
How to Define a Discrete Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1278
Specifying the Attribute Mapping Information in a
DISCRETEATTRMAP Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1278
Specifying the Attribute Mapping Information in a SAS Data Set . . . . . . . . . . . 1279
Defining a Range Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1281
Referencing an Attribute Map in Your Plot Statements . . . . . . . . . . . . . . . . . . . . 1282
Referencing a Discrete Attribute Map in Plot Statements . . . . . . . . . . . . . . . . . . 1282
Referencing a Range Attribute Map in Plot Statements . . . . . . . . . . . . . . . . . . . . 1285

About Attribute Maps

By default, many of the graphical attributes of a plot vary with the plot data. For
example, when plots display grouped values, by default, the graphical attributes for each
group value are selected from the GraphData1–GraphDataN style elements in data order.
Changes in the data order can significantly change the appearance of the plot. When
plots display a color gradient, by default, the colors assigned to the classification
variable are derived from a color ramp based on the actual range of the data. The color
assigned to each value can vary with the range of the classification values. Attribute
maps enable you to assign the same graphical properties to specific values or ranges of
values regardless of data order or the data range. They are useful when you want your
graphs to have consistent visual properties when the data varies.
The GTL supports two types of attribute maps: discrete attribute maps and range
attribute maps. A discrete attribute map maps discrete values to graphical properties. A
range attribute map maps numeric values or ranges of numeric values to graphical
properties.
1278 Chapter 17 • Key Concepts for Using Attribute Maps

Defining a Discrete Attribute Map

How to Define a Discrete Attribute Map

You define a discrete attribute map in the following way:
• Specify the attribute mapping information in a SAS data set or in a
DISCRETEATTRMAP block in the template code. See “Specifying the Attribute
Mapping Information in a DISCRETEATTRMAP Block” on page 1278 and
“Specifying the Attribute Mapping Information in a SAS Data Set” on page 1279.
Note: Defining a discrete attribute map in a SAS data set is valid in the first
maintenance release of SAS 9.4 and later releases.
• Reference the attribute map in plot statements by using one of the following
methods:
• Include a DISCRETEATTRVAR statement in the template code to create an
attribute map variable. The variable associates the attribute map with a
classification column in the plot data. Then use the attribute map variable to
reference the discrete attribute map in plot statements. This method applies to an
attribute map that is defined in a SAS data set or in a DISCRETEATTRMAP
block in the template code.
• Specify the classification column, which is in the plot data, in the plot statements
where needed. When the graph is rendered, the SGRENDER DATTRVAR
statement is used to associate the attribute map that is defined in a SAS data set
with the classification column in the plot data. This method applies only to an
attribute map that is defined in a SAS data set.
Note: The DATTRVAR statement is valid in the first maintenance release of
SAS 9.4 and later releases. For information about the DATTRVAR statement,
see SAS ODS Graphics: Procedures Guide.

Specifying the Attribute Mapping Information in a

DISCRETEATTRMAP Block
Use a DISCRETEATTRMAP block to specify the attribute mapping information in the
template code. Place the block in the global definition area of your template between the
BEGINGRAPH statement and the first layout statement. The block contains one or more
VALUE statements that specify the attribute mapping information. Each VALUE
statement specifies a single discrete value and one or more graphical properties that are
assigned to that value. The NAME= option in the DISCRETEATTRMAP statement
specifies the name of the attribute map. The DISCRETEATTRVAR statement creates the
reference variable for the attribute map. The following options are in the
DISCRETEATTRVAR statement:
• The ATTRVAR= option specifies a unique name for the attribute-map-to-data-set-
column association.
• The ATTRMAP= option specifies the value of the NAME= option that is specified in
the DISCRETEATTRMAP statement.
• The VAR= option specifies the name of the numeric or character column in the plot
data set, an expression, or the name of a dynamic variable.
Defining a Discrete Attribute Map 1279

Use the attribute map variable to reference the attribute map in plot statements.
Note: Do not use the attribute variable in an expression. Doing so might produce
unexpected results.
Note: The values and graphical attributes that are defined in a discrete attribute map
cannot be displayed by a CONTINUOUSLEGEND statement.
For for more information, see “Defining a Discrete Attribute Map in a
DISCRETEATTRMAP Block” on page 1292.

Specifying the Attribute Mapping Information in a SAS Data Set

Starting with the first maintenance release of SAS 9.4, you can specify the attribute
mapping information for a discrete attribute map in a SAS data set instead of in a
DISCRETEATTRMAP block in the template. Each observation in the data set maps a
discrete value to one or more specific graphical properties. The following table lists the
columns that you must include in your data set for each observation.

Table 17.1 Columns That Are Required in Each Observation

Column Name Type Description

ID Character Specifies the name of the attribute map to which this

observation is associated. The ID value must be a SAS
name token and not a literal. A data set can define
multiple attribute maps. All observations that are
associated with a specific attribute map must have the
same ID value. When you are creating an attribute map
variable for this attribute map, specify the name that you
specified in the ID column in the DISCRETEATTRVAR
statement ATTRMAP= argument. The ID column can
also be specified in an SGRENDER DATTRVAR
statement when the graph is rendered. The ID value is
case sensitive.

Value Character Specifies the discrete classification value that is to be

mapped to the graphical properties that are specified in
this observation. The value is case sensitive.

The following table lists the columns that you can use to specify the graphical properties
for each classification value. Properties that you do not specify default to the properties
that are normally used when an attribute map is not specified.

Table 17.2 Optional Columns That Specify Graphical Properties

Column Name Type Description

FillColor Character Specifies the color for the filled areas for this
classification value. When FillStyleElement is also
specified, this column overrides the specified style
element’s color attribute.

FillStyleElement Character Specifies the name of a style element, such as

GraphData3, that is to provide the fill attributes.
1280 Chapter 17 • Key Concepts for Using Attribute Maps

Column Name Type Description

FillTransparency Numeric Specifies the transparency of the fill color for this
classification value. Values are in the range 0 (opaque)
to 1 (completely transparent).

LineColor Character Specifies the line color for this classification value.
When LineStyleElement is also specified, this column
overrides the specified style element’s contrast color
attribute.

LinePattern Character Specifies the line pattern for this classification value.
or Numeric You can specify the pattern name or number. See
“Available Line Patterns” on page 1352.

LineStyleElement Character Specifies the name of a style element, such as

GraphData3, that is to provide the line attributes.

LineThickness Numeric Specifies the line thickness, in pixels, for this

classification value. The values must be integers.

MarkerColor Character Specifies the color of the marker symbol for this
classification value. When MarkerStyleElement is also
specified, this column overrides the specified style
element’s contrast color attribute.

MarkerSize Numeric Specifies the size of the marker, in pixels, for this
classification value. The values must be integers.

MarkerSyleElement Character Specifies the name of a style element, such as

GraphData3, that is to provide the marker attributes.

MarkerSymbol Character Specifies the marker symbol to use for this classification
value. See SYMBOL= in “Marker Options” on page
1350. When MarkerStyleElement is also specified, this
column overrides the specified style element’s marker
symbol attribute.

MarkerTransparency Numeric Specifies the transparency of the marker for this

classification value. Values are in the range 0 (opaque)
to 1 (completely transparent).

NoCase Character Specifies whether the attribute-map value comparisons

are case-insensitive. Valid values are TRUE and
FALSE. When TRUE, value comparisons are case
insensitive. The default is TRUE.
Note: This column is valid starting with the third
maintenance release of SAS 9.4.

TextColor Character Specifies the text color for this classification value.

TextFont Character Specifies the text font by name for this classification
value.
Defining a Range Attribute Map 1281

Column Name Type Description

TextSize Numeric Specifies the text font size in points for this
classification value. The values must be integers.

TextStyle Character Specifies the text style for this classification value. Valid
values are NORMAL and ITALIC.

TextStyleElement Character Specifies the name of a style element, such as

GraphLabelText, that is to provide the text attributes.

TextWeight Character Specifies the text weight for this classification value.
Valid values are NORMAL and BOLD.

Show Character Specifies whether the legend displays only the attribute
map values that appear in the data or always displays all
of the values in the attribute map. Valid values are
DATA and ATTRMAP. The default is DATA.
Note: This column is valid starting with the third
maintenance release of SAS 9.4.

The ID column provides the name of the attribute map, which is specified in the
ATTRMAP= option in the DISCRETEATTRVAR statement or in a DATTRVAR
statement. When the graph is rendered, the name of the attribute map data set must be
specified in the DATTRMAP= option in the SGRENDER statement. For information
about how to reference the attribute map in your plot statements, see “Referencing a
Discrete Attribute Map in Plot Statements” on page 1282.

Defining a Range Attribute Map

A range attribute map is defined as follows:
• The attribute mapping information is specified in a RANGEATTRMAP block in the
template code. It must be placed in the global definition area of your template
between the BEGINGRAPH statement and the first layout statement. The NAME=
option in the RANGEATTRMAP statement specifies the name of the attribute map.
The block contains one or more RANGE statements that specify the attribute
mapping information. Each RANGE statement specifies a numeric value or numeric
value range, and one or more graphical properties that are assigned to that value or
range. An ENDRANGEATTRMAP statement is used to close the block.
Note: Range attribute maps do not support attribute map information that is stored in
a SAS data set.
• The RANGEATTRVAR statement is used to create an attribute map variable that
associates the attribute map with a column in the data that provides the classification
values. In the RANGEATTRVAR statement:
• The ATTRVAR= option specifies a unique name for range attribute map to data
set column association.
• The ATTRMAP= option specifies the value of the NAME= option that is
specified in the RANGEATTRMAP statement.
1282 Chapter 17 • Key Concepts for Using Attribute Maps

• The VAR= option specifies the name of the numeric column in the plot data set
with which the range attribute map is to be associated.
Note: A RANGEATTRMAP can be used with a numeric column only.
• The attribute map variable is used to reference the discrete attribute map in plot
statements.
Note: The values and graphical attributes defined in a range attribute map cannot be
displayed by a DISCRETELEGEND statement.
For information about how to reference a range attribute map, see “Referencing a
Discrete Attribute Map in Plot Statements” on page 1282.
The RANGE statements in the RANGEATTRMAP block can associate a range of values
or a single value with a single color or a color ramp. The syntax of the RANGE
statement is as follows:
RANGE low-value< < > – < < >high-value / options
The optional exclusion operator (<) can be placed after the low value or before the high
value to exclude that value from the range endpoint. The low value and high value can
be an unformatted numeric value or a range keyword. For the low value, keyword MIN,
NEGMAX, or NEGMAXABS can be used instead of numeric value. For the high value,
keyword MAX or MAXABS can be used. For information about the range keywords,
see SAS Graph Template Language: Reference.
Note: If two ranges share a common endpoint, such as 10–20 and 20–30, and no
exclusion operator ( < ) is used, then the common endpoint belongs to the lower
range, which is 10–20 in this case.
For for more information about how to create a range attribute map, see “Creating and
Using a Range Attribute Map” on page 1305.

Referencing an Attribute Map in Your Plot

Statements

Referencing a Discrete Attribute Map in Plot Statements

If you use a DISCRETEATTRVAR statement to create a variable for an attribute map,
then reference the attribute map in a plot statement by specifying the name that is
assigned in the DISCRETEATTRVAR statement’s ATTRVAR= argument. Each plot
statement’s documentation in this reference indicates which options support a reference
to a discrete attribute map variable as the specified value.
Note: A reference to a discrete attribute variable must be a direct reference to the
attribute variable. It cannot be set by a dynamic variable.
When you use a DATTRVAR statement in the SGRENDER statement, the attribute map
is referenced implicitly when a column that is assigned to the attribute map ID in the
DATTRVAR statement is referenced in a plot statement. The following table lists the
statement options that support a reference to a discrete attribute map.
Referencing an Attribute Map in Your Plot Statements 1283

Table 17.3 Statement Options That Support a Discrete Attribute Map

Statement Options

AXISTABLE COLORGROUP= on page 193

TEXTGROUP= on page 200

BANDPLOT GROUP= on page 211

BARCHART GROUP= on page 232

BARCHARPARM GROUP= on page 267

BOXPLOT GROUP= on page 317

BOXPLOTPARM GROUP= on page 350

BUBBLEPLOT GROUP= on page 377

ELLIPSEPARM GROUP= on page 434

FRINGEPLOT GROUP= on page 441

HEATMAPPARM COLORGROUP= on page 461

HIGHLOWPLOT GROUP= on page 480

LINECHART GROUP= on page 480

LINEPARM GROUP= on page 549

LOESSPLOT GROUP= on page 559

MOSAICPLOTPARM COLORGROUP= on page 575

NEEDLEPLOT GROUP= on page 592

PBSPLINEPLOT GROUP= on page 606

PIECHART CATEGORY= on page 615

POLYGONPLOT GROUP= on page 636

REGRESSIONPLOT GROUP= on page 673

SCATTERPLOT GROUP= on page 694

SCATTERPLOTMATRIX GROUP= on page 727

SERIESPLOT GROUP= on page 758

STEPPLOT GROUP= on page 791

1284 Chapter 17 • Key Concepts for Using Attribute Maps

Statement Options

VECTORPLOT GROUP= on page 846

WATERFALLCHART COLORGROUP= on page 860

In a DISCRETELEGEND statement, reference the plot statement that uses the attribute
map. The plot statement must have a NAME= option that assigns a name to the plot,
because the DISCRETELEGEND statement references that name. Because the attribute
map is referenced in the plot statement, the legend uses the attribute map to represent the
group values that exist in the data.
TIP If the discrete attribute map is defined in a DISCRETEATTRMAP block, then
you can use the DISCRETELEGENDENTRYPOLICY=ATTRMAP option in the
DISCRETEATTRMAP statement to display all of the items that are defined in the
attribute map regardless of whether the values appear in the data. See
DISCRETELEGENDENTRYPOLICY= on page 1288.
For an example of how to reference an attribute map variable that is created with the
DISCRETEATTRVAR statement, see “Example: DISCRETEATTRVAR Statement with
an Attribute Map Data Set” on page 1298.
The following code shows you how to reference an attribute map when a DATTRVAR
statement is used in the SGRENDER statement to associate discrete attribute map
SYMBOLS with classification column Sex. The SYMBOLS attribute map is defined in
a SAS data set, and the column Sex is referenced in a SCATTERPLOT statement.
/* Create the attribute map data set */
data attrds;
input ID $1-7 VALUE $9 MARKERSYMBOL $11-23 MARKERCOLOR $25-30;
datalines;
symbols M diamondfilled blue
symbols F circlefilled red
;
run;

/* Define the template for this graph */

proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Height and Weight by Sex";
layout overlay;
scatterplot x=height y=weight / name="scatter"
group=sex;
discretelegend "scatter";
endlayout;
endgraph;
end;

/* Generate the graph */

proc sgrender data=sashelp.class dattrmap=attrds template=scatterplot;
dattrvar sex="symbols";
run;

The output is shown in “Example: DISCRETEATTRVAR Statement with an Attribute

Map Data Set” on page 1298.
Referencing an Attribute Map in Your Plot Statements 1285

Referencing a Range Attribute Map in Plot Statements

In a plot statement, you reference a range attribute map by specifying the name that is
assigned in the RANGEATTRVAR statement’s ATTRVAR= argument. Each plot
statement’s documentation in this reference indicates which options support a reference
to a range attribute map variable as the specified value.
Note: A reference to a range attribute variable must be a direct reference to the attribute
variable. It cannot be set by a dynamic variable.
The following table lists the statement options that support a reference to a range
attribute map variable.

Table 17.4 Statement Options That Support a Range Attribute Map Variable Reference

Statement Options

BUBBLEPLOT COLORRESPONSE=

HEATMAP COLORRESPONSE=

HEATMAPPARM COLORRESPONSE=

MOSAICPLOTPARM COLORRESPONSE=

POLYGONPLOT COLORRESPONSE=

SCATTERPLOT COLORRESPONSE= or MARKERCOLORGRADIENT=

SCATTERPLOTMATRIX COLORRESPONSE= or MARKERCOLORGRADIENT=

SURFACEPLOTPARM COLORRESPONSE or SURFACECOLORGRADIENT=

WATERFALLCHART COLORRESPONSE=

In a CONTINUOUSLEGEND statement, reference the plot statement that uses the

attribute map. The plot statement must have a NAME= option that assigns a name to the
plot, because the CONTINUOUSLEGEND statement references that name. Because the
attribute map is referenced in the plot statement, the legend uses that attribute map to
represent the numeric values that are present in the data.
For an example, see “Example: RANGEATTRMAP and RANGEATTRVAR
Statements” on page 1307.
1286 Chapter 17 • Key Concepts for Using Attribute Maps
1287

Chapter 18
Discrete Attribute Map
Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1287
DISCRETEATTRMAP Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1287
DISCRETEATTRVAR Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1297

Dictionary

DISCRETEATTRMAP Statement
Defines a set of graphical properties that can be associated with user-defined sets of values.
Restriction: The DISCRETEATTRMAP block cannot be nested within any other block.
Requirements: The DISCRETEATTRMAP block and the DISCRETEATTRVAR statement must
appear in the global definition area of the template between the BEGINGRAPH
statement and the first LAYOUT statement.
The DISCRETEATTRMAP block must contain at lease one VALUE statement.
The DISCRETEATTRVAR statement must be used to associate the discrete attribute
map with a data column.
Notes: The graphical properties for a discrete attribute map can also be defined in a SAS
data set. See “Defining a Discrete Attribute Map” on page 1278.
Prior to the third release of SAS 9.4, when a discrete attribute map is used for group
values in a plot that contributes to a discrete legend and attributes are overridden in
the plot statement, the attributes of some plot features and their corresponding
legend items might not match. Starting with the third maintenance release of SAS
9.4, the attributes of the legend items always match the attributes of the
corresponding plot features.
See: “DISCRETEATTRVAR Statement” on page 1297

Syntax
DISCRETEATTRMAP NAME="string" </option(s)>;
VALUE value-spec </option(s)>;
<… more VALUE statements …>
ENDDISCRETEATTRMAP;
1288 Chapter 18 • Discrete Attribute Map Statements

Required Argument
NAME="string "
assigns a name to the attribute definition. The name can be referenced in a
DISCRETEATTRVAR statement, which is used to associate the attribute map with
an input data column. The name can also be referenced in a DISCRETELEGEND
statement to map the specified graphical properties directly to a discrete legend.

Restriction The string is case sensitive, must be enclosed in quotation marks, and
must define a unique name within the template.

Optional Arguments
DISCRETELEGENDENTRYPOLICY=DATA | ATTRMAP
specifies whether the items that are contributed to a discrete legend by the plot
associated with this attribute map are only items that appear in the data or only items
that are defined in the attribute map.
Note: This option is valid in the first maintenance release of SAS 9.4 and later
releases.
DATA
the associated plot contributes to a discrete legend only items that appear in the
data.
ATTRMAP
the associated plot contributes to a discrete legend only items that are defined in
the discrete attribute map.

Interaction If this option is set to ATTRMAP, then data skins, overrides from
the DATATRANSPARENCY= option, and overrides from the
TRANSPARENCY= suboption in the FILLATTRS= and
MARKERATTRS= options are displayed in the discrete legend.
Overrides from other options such as the COLOR= suboption in the
FILLATTRS= option are not displayed.

Default DATA

IGNORECASE=TRUE | FALSE
specifies whether case is ignored when comparing the values that are specified in the
attribute map with values from an input data column.

Default FALSE. Value comparisons are case-sensitive.

Tip The effect of this option can be achieved by applying a function like
UPCASE to the data column and using only uppercase strings in each
VALUE statement.

See “boolean ” on page 1339 for other Boolean values that you can use.

TRIMLEADING=TRUE | FALSE
specifies whether leading blanks are trimmed (removed) from both the attribute map
values and the input data values before those values are compared. Trailing blanks
are always trimmed.

Default TRUE. Leading blanks are trimmed.

See “boolean ” on page 1339 for other Boolean values that you can use.
DISCRETEATTRMAP Statement 1289

VALUE Statement Required Argument

value-spec
specifies one or more formatted strings or the keyword OTHER. Strings are always
quoted. Multiple strings must be separated by blanks, and each of the strings must be
enclosed in its own set of quotation marks. The formatted strings must be equal to
the formatted values of the classification column that is used with the
DISCRETEATTRVAR statement.
OTHER
creates a category for all other column values that are not explicitly assigned with
VALUE statements. This keyword is not quoted. The default attributes for these
values are derived from the GraphOther style element.

Note If OTHER is not specified, then data values that are not explicitly
assigned with VALUE statements are mapped to attributes as if a discrete
attribute map is not in effect.

The following examples elaborate on the value-spec strings:

"Hybrid" By default, all string comparisons are case-sensitive. By

default, the string Hybrid does not match the string HYBRID.
"HYBRID" If IGNORECASE=TRUE in the DISCRETEATTRMAP
statement, then you can specify an upper-, lower-, or mixed-
cased string for the value-spec string. When
IGNORECASE=TRUE, the string Hybrid matches the string
HYBRID.
"15JAN2011" If a numeric column is being mapped with a VALUE
statement, then you must specify the formatted value of the
column. This example shows how to specify the value of a
numeric SAS date column that has a DATE9. format
associated with it.
"." If a numeric column has a missing value, then you should use
the formatted value for missing, which is "." by default. If the
MISSING= system option is used to change the default string,
then you should match that value. For example, if OPTIONS
MISSING="M" is specified in the SAS program, then you
should use "M" in the VALUE statement to represent missing
values.
"" If a character column has a missing value, then you should use
the formatted value for missing, which is " " by default.
"Truck" Multiple strings can be specified to indicate that each of the
"SUV" specified values matches to the same graphical properties. It
does not mean that a single new category is formed. The list
of strings is separated by blanks, and each string is enclosed in
its own set of quotation marks.

Restriction If a user-defined format is associated with the classification column,

then you should specify the same formatted strings that appear in the
format definition.

Note When the specified value strings are compared with the data values,
leading blanks are honored and trailing blanks are ignored.
1290 Chapter 18 • Discrete Attribute Map Statements

Tip In the first maintenance release of SAS 9.4 and earlier releases, if you
create and use a format to display a special value for missing character
values, in some cases, " " is returned instead of the formatted value. If
your attribute map assigns attributes to the formatted missing value in
that case, the attributes are not assigned to the missing values. To
correct this problem, specify both " " and your formatted missing value
in the VALUE statement for your missing-value attributes. This issue is
resolved staring with the second maintenance release of SAS 9.4.

VALUE Statement Optional Arguments

The following options can be used in the VALUE statement.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
the fill attributes to be used when an attribute map is applied to filled areas in a
graph.

Defaults When DISCRETELEGENDENTRYPOLICY=DATA, unspecified

attributes receive the attributes that they would have if the attribute
map were not defined.

When DISCRETELEGENDENTRYPOLICY=ATTRMAP, unspecified

attributes derive attributes from the GraphDataDefault style element.

Restriction In SAS 9.4 and earlier releases, the TRANSPARENCY= fill option is
ignored. Starting with the first maintenance release of SAS 9.4, the
TRANSPARENCY= fill option is supported.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Fill Options” on page 1348 for available fill-options.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes to be used when an attribute map is applied to lines in a
graph.

Defaults When DISCRETELEGENDENTRYPOLICY=DATA, unspecified

attributes receive the attributes that they would have if the attribute
map were not defined.

When DISCRETELEGENDENTRYPOLICY=ATTRMAP, unspecified

attributes derive attributes from the GraphDataDefault style element.

Restriction In SAS 9.4 and earlier releases, the THICKNESS= line option is
ignored. Starting with the first maintenance release of SAS 9.4, the
THICKNESS= line option is supported.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style-element.

“Line Options” on page 1349 for available line-options.

MARKERATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the marker attributes to be used when an attribute map is applied to marker
symbols in a graph.
DISCRETEATTRMAP Statement 1291

Defaults When DISCRETELEGENDENTRYPOLICY=DATA, unspecified

attributes receive the attributes that they would have if the attribute
map were not defined.

When DISCRETELEGENDENTRYPOLICY=ATTRMAP,
unspecified attributes derive attributes from the GraphDataDefault
style element.

Restrictions The WEIGHT= marker option is ignored.

In SAS 9.4 and earlier releases, the SIZE=, TRANSPARENCY=, and

WEIGHT= marker options are ignored. Starting with the first
maintenance release of SAS 9.4, the SIZE=, TRANSPARENCY=, and
WEIGHT= marker options are supported.

See “General Syntax for Attribute Options” on page 1347 for the syntax
on using a style-element.

“Marker Options” on page 1350 for available marker-options.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text attributes to use when an attribute map is applied to text in a graph.

Default The GraphDataText style element.

Restriction This option is honored only by the TEXTGROUP=option in the

AXISTABLE statement and the GROUP= option in the TEXTPLOT
statement.

See “General Syntax for Attribute Options” on page 1347 for the syntax on
using a style element.

“Text Options” on page 1351 for available text options.

Details

About Discrete Attribute Maps

Attribute maps can be useful for ensuring that a particular value (a company name, for
example) is always represented by the same visual characteristics in your graphs,
regardless of the value’s order in the input data. When specified directly in a discrete
legend, an attribute map can be used to display legend entries for group values that are
not in the data. (See “Displaying Legend Entries for Group Values That Are Not in the
Data” on page 1293.)
If you do not want to manage the graphical properties that are associated with each
unique discrete value, then you can simply specify an input column in the grouping
option. In that case, each discrete value is represented by different graphical properties in
the graph. The default properties are derived from options that are set in the plot
statement or from the GraphData1–GraphDataN style elements that are defined in the
ODS style that is in effect. However, to ensure that specific graphical properties are used
to represent a discrete value in the graph, regardless of that value’s order in the data, you
can use an attribute map to create that association. Any values in the data that are not
accounted for in the attribute map are assigned the graphical properties that they would
receive if the attribute map is not defined.
1292 Chapter 18 • Discrete Attribute Map Statements

Defining a Discrete Attribute Map in a DISCRETEATTRMAP Block

A DISCRETEATTRMAP block creates an attribute map that matches graphical
properties to discrete values. The attribute map can be associated with a data input
column that is used as a classification variable in a graph. It can also be specified
directly in a discrete legend.
To define a discrete attribute map in a DISCRETEATTRMAP block, do the following:
• Use the DISCRETEATTRMAP statement to start the attribute definition and assign a
name to it. The DISCRETEATTRMAP statement must be located within the
BEGINGRAPH block but outside of the outermost layout block. It cannot be nested
in another DISCRETEATTRMAP statement or in a RANGEATTRMAP statement.
The DISCRETEATTRMAP statement determines whether the data mapping is case-
sensitive and whether leading blanks are trimmed from the data values during the
mapping.
• Nest within the DISCRETEATTRMAP block at least one VALUE statement that
specifies graphical properties to associate with a classification value. See “About the
VALUE Statement” on page 1292. Use one VALUE statement for each
classification value that you want to map. Values that are referenced in the attribute
map can be character or numeric. Discrete values that are not accounted for in the
attribute map are assigned the graphical properties that they would receive if the
attribute map is not defined.
• Use the ENDDISCRETEATTRMAP statement to close the block.
For information about how to use the attribute map, see “Using a Discrete Attribute
Map” on page 1293.

About the VALUE Statement

A VALUE statement within the DISCRETEATTRMAP block associates graphical
properties with a discrete value in the attribute map. To associate graphical properties
with multiple values, specify multiple VALUE statements in the attribute map using the
following general syntax for each of the statements:
VALUE value-spec </option(s)>;
For a specific example, see the “Example Program” on page 1296.
If the discrete attribute map is referenced by a plot statement in the template, then the
graphical properties that are defined in the VALUE statements are used in the plot. If a
discrete legend is generated for the plot, then the graphical properties are represented in
that legend.
If a discrete attribute map is referenced directly in a DISCRETELEGEND statement,
then the graphical properties that are defined in the VALUE statement are mapped
directly to the legend and are independent of the values in the data. For more
information and an example, see the “Defining a Discrete Attribute Map in a
DISCRETEATTRMAP Block” on page 1292 .
If two or more VALUE statements define attributes to associate with the same
classification values, then the last VALUE statement's settings are used.
By default when comparing a column's value to a string that is specified for the VALUE
statement's value-spec,
• the column value is formatted to a string, using the format that is defined for the
column or the default format if no format is defined for the column
• leading spaces are trimmed from the string that is specified in the VALUE statement
DISCRETEATTRMAP Statement 1293

• a case-sensitive comparison is performed between the column string and the VALUE
string.
To change the default behavior for the comparison, you can use the
DISCRETEATTRMAP statement’s TRIMLEADING= and IGNORECASE= options.

Using a Discrete Attribute Map

To use a discrete attribute map that is defined in a DISCRETEATTRMAP block, do the
following:
• Use the DISCRETEATTRVAR statement to create a named association between the
defined attribute map and the input column that contains the classification values.
The ATTRMAP= argument identifies the attribute map and the VAR= argument
identifies the input column. Use the ATTRVAR= argument to assign a name that can
be used to reference the named association in plot statements within the template.
• Reference the attribute map variable in the plot statements where needed. See
“Referencing a Discrete Attribute Map in Plot Statements” on page 1282.
For an example, see “Example Program” on page 1296.

Displaying Legend Entries for Group Values That Are Not in the
Data
To display legend entries for the values in a grouped plot, you typically use the plot
statement’s NAME= option to assign a name to the plot, and then reference that name in
the DISCRETELEGEND statement. In this usage case, the legend displays entries for
the group values that exist in the data.
To represent all of the group values in the legend, regardless of whether they exist in the
data, you can specify DISCRETELEGENDENTRYPOLICY=ATTRMAP in the
DISCRETEATTRMAP statement for the attribute map. When the
DISCRETELEGENDENTRYPOLICY=ATTRMAP option is in effect, the associated
plot contributes all of the items in the attribute map to the discrete legend regardless of
whether they exist in the data.
Displaying all of the items in a discrete attribute map in the legend can be useful for
flagging data in the graph. For example, assume you have weight and height values for
all students in an analysis group and you want to create a scatter plot of the data.
However, some of the observations are incomplete and do not record the student’s sex.
You want to include the incomplete observations in your plot, but you want to visually
distinguish them from the others. In that case, you can do the following:
• Represent the unknown values in your data by entering the value U for sex.
• Define the discrete attribute map for the plot in a DISCRETEATTRMAP block.
Include the DISCRETELEGENDENTRYPOLICY=ATTRMAP option in the
DISCRETEATTRMAP statement. Include VALUE statements that specify the
properties for the M, F, and U values.
• Use the DISCRETEATTRVAR statement to create a discrete attribute map variable
that associates the attribute map with column Sex in the input data.
• In the SCATTERPLOT statement, specify the discrete attribute map variable name in
the GROUP= option.
• In the DISCRETELEGEND statement, reference the scatter plot.
As a result, the legend displays the attribute-map definitions, and observations with the
value U in column Sex are displayed as incomplete observations.
1294 Chapter 18 • Discrete Attribute Map Statements

Here is an example of a template that uses a discrete attribute map to uniquely display
observations with the value U in column Sex.
proc template;
define statgraph discreteattrmapdatapresent;
begingraph;
entrytitle "Height and Weight by Sex";
/* Define the attribute map and assign the name "symbols" */
discreteattrmap name="symbols" / trimleading=true ignorecase=true
discretelegendentrypolicy=attrmap;
value "M" / markerattrs=(color=blue symbol=diamondfilled);
value "F" / markerattrs=(color=green symbol=circlefilled);
value "U" / markerattrs=(color=red symbol=starfilled);
enddiscreteattrmap;

/* Create attribute map variable GROUPMARKERS to associate attribute

map SYMBOLS with column Sex */
discreteattrvar attrvar=groupmarkers var=sex attrmap="symbols";

/* Use the attribute map by referencing GROUPMARKERS in a plot

statement */
layout overlay;
scatterplot x=height y=weight / name="scatter"
group=groupmarkers;
discretelegend "scatter";
endlayout;
endgraph;
end;

To test the template, you can generate test data set Testclass from Sashelp.Class by
changing the value of column Sex to U for John and Carol as shown in the following
code.
data testclass;
set sashelp.class;
if (name="John") then sex="U";
if (name="Carol") then sex="U";
run;

The following figure shows the output of this template when it is run with the test data
set Testclass.

The two red stars in the plot indicate the observations with the value U in column Sex.
The next figure shows the output of this template when it is run with the complete data
in data set Sashelp.Class.
Example: DISCRETEATTRMAP and DISCRETEATTRVAR Statements 1295

The absence of the red stars in the plot indicates that the value U is no longer present in
the data. All of the observations are now complete. When
DISCRETELEGENDENTRYPOLICY=ATTRMAP is specified for an attribute map, be
aware that the legend entries that are contributed by an associated plot are defined
entirely by the attribute map and are independent of the data.

Example: DISCRETEATTRMAP and

DISCRETEATTRVAR Statements

The following graph was generated by the “Example Program” on page 1296. The
example defines graphical properties to associate with classification values in an input
column that is used in a scatter plot. The DISCRETEATTRMAP statement starts the
attribute map definition, assigns a name to it, and ensures that the data mapping is not
case sensitive. The VALUE statements define the colors and marker symbols to associate
with the values M and F. The DISCRETEATTRVAR statement associates the attribute
map with the data column Sex and assigns the name GROUPMARKERS to the
association. The SCATTERPLOT statement references the named association in its
GROUP= option.
1296 Chapter 18 • Discrete Attribute Map Statements

Example Program
proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Height and Weight by Sex";

/* define the attribute map and assign the name "symbols" */

discreteattrmap name="symbols" / ignorecase=true ;
value "m" / markerattrs=(color=blue symbol=diamondfilled) ;
value "f" / markerattrs=(color=red symbol=circlefilled) ;
enddiscreteattrmap ;

/* associate the attribute map with input data column Sex and assign
* the name GROUPMARKERS to the named association */
discreteattrvar attrvar=groupmarkers var=sex attrmap="symbols" ;

/* reference GROUPMARKERS in a plot statement */

layout overlay;
scatterplot x=height y=weight / name="scatter"
group=groupmarkers ;
discretelegend "scatter";
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=scatterplot;

run;
DISCRETEATTRVAR Statement 1297

DISCRETEATTRVAR Statement
Creates a named association between a user-defined discrete attribute map and an input data column.
Restriction: The DISCRETEATTRVAR statement cannot appear within a DISCRETEATTRMAP
or RANGEATTRMAP block.
Requirement: A discrete attribute map must be created using the DISCRETEATTRMAP statement.
See: “Example: DISCRETEATTRMAP and DISCRETEATTRVAR Statements” for an
example.

Syntax
DISCRETEATTRVAR ATTRVAR=attrvar-name
VAR=data-column | expression | dynamic
ATTRMAP="attrmap-name";

Required Arguments
ATTRVAR=attrvar-name
specifies a SAS name for this association between the attribute map and the input
column. This name must be unique within the template and can be referenced by
other statements that can be associated with the attribute map. The attribute map
variable name should not be used in an expression. If it is, then the results are
unpredictable.

Restriction The name that is assigned in this argument is used to associate an

attribute map with the discrete values in an input data column. Thus, it
is not the name to reference when you want to display legend entries
that are independent of the data. For that special use, a
DISCRETELEGEND statement can reference the attribute map
directly by the name that is assigned in the DISCRETEATTRMAP
statement. For more information, see the DISCRETEATTRMAP
statement’s “Defining a Discrete Attribute Map in a
DISCRETEATTRMAP Block” on page 1292 .

Note The assigned SAS name can be the same as the name of the data input
column, but it is not recommended. If an assigned attrvar-name
matches the name of an input data column, then the attrvar-name takes
precedence.

VAR=data-column | expression | dynamic

specifies an input data column to be associated with an attribute map at run time. If
an expression is used, a new column of transformed values is created and then
matched with the attribute map.

Interaction If the column is not found or the column is of the wrong type for the
attribute map, then the DISCRETEATTRVAR statement is ignored.

Tip The input data column can be character or numeric, but the values must
match the type of the values that are specified in the attribute map. For
numeric columns, all values are treated as discrete values.
1298 Chapter 18 • Discrete Attribute Map Statements

ATTRMAP="attrmap-name "
specifies the name of an existing discrete attribute map.

Restriction The attrmap-name is case sensitive, must be enclosed in quotation

marks, and must be the name that was assigned to the attribute map in
the DISCRETEATTRMAP statement’s NAME= argument.

Details
The DISCRETEATTRVAR statement creates and names an association between
graphical properties that are specified in a DISCRETEATTRMAP block and a
classification column that is in the data. The name that is assigned to the association in
the DISCRETEATTRVAR statement is the name that plot statements must reference to
use the attribute map.
Defining and using a discrete attribute map requires you to coordinate settings on several
statements. For more information, see the DISCRETEATTRMAP statement’s “Defining
a Discrete Attribute Map in a DISCRETEATTRMAP Block” on page 1292 .
The DISCRETEATTRVAR statement must be located within the BEGINGRAPH block
but outside of the outermost layout block. It cannot be nested in a
DISCRETEATTRMAP statement.

Example: DISCRETEATTRVAR Statement with an

Attribute Map Data Set

Starting with the first maintenance release of SAS 9.4, you can define an attribute map in
a SAS data set. This example shows you how to use the DISCRETEATTRVAR
statement with an attribute map that is defined in a SAS data set. It is the example in
“Example: DISCRETEATTRMAP and DISCRETEATTRVAR Statements” on page
1295 modified to use a SAS data set instead of a DISCRETEATTRMAP block.
Example: DISCRETEATTRVAR Statement with an Attribute Map Data Set 1299

Here is the output for this example.

Example Program
Here is the SAS code for this example.
/* Create the attribute map data set */
data attrds;
input ID $1-7 VALUE $9 MARKERSYMBOL $11-23 MARKERCOLOR $25-30;
datalines;
symbols M diamondfilled blue
symbols F circlefilled red
;
run;

/* Define the template for this graph */

proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Height and Weight by Sex";

/* Associate the attribute map with input data column Sex and assign
* the name GROUPMARKERS to the named association */
discreteattrvar attrvar=groupmarkers var=sex attrmap="symbols";

/* Reference GROUPMARKERS in a plot statement */

layout overlay;
scatterplot x=height y=weight / name="scatter"
group=groupmarkers;
discretelegend "scatter";
endlayout;
endgraph;
end;
1300 Chapter 18 • Discrete Attribute Map Statements

/* Generate the graph */

proc sgrender data=sashelp.class dattrmap=attrds template=scatterplot;
run;

Details
The attribute map is defined in data set Attrds. The data set includes the required
columns ID and VALUE, which specify a name for this attribute map and the value of
the classification variable that is to be mapped. The MarkerSymbol and MarkerColor
columns specify the graphical properties that are mapped to each value. The values
specified in the Value column are case-sensitive. Unlike the DISCRETEATTRMAP
statement, when you define your attribute map in a data set, there is no option that you
can specify to ignore case. You must ensure that the case of the values in your attribute
map matches the case in the actual data.
The DISCRETEATTRVAR statement associates the attribute map with the data column
Sex and assigns the name GROUPMARKERS to the attribute map variable. The
ATTRMAP= option specifies the name in the ID column of the attribute map data set.
The value is case-sensitive. The attribute map variable name is specified in the GROUP=
option in the SCATTERPLOT statement.
Finally, in the SGRENDER statement, the DATTRMAP= option specifies the name of
the attribute map data set. When your attribute map is defined in a SAS data set, you
must specify the name of the data set in the DATTRMAP= option in the SGRENDER
statement that renders the graph.
You can also use the DATTRVAR= statement with the SGRENDER statement to
associate the attribute map with the data column Sex. See “Referencing an Attribute
Map in Your Plot Statements” on page 1282.
1301

Chapter 19
Range Attribute Map Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1301
RANGEATTRMAP Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1301
RANGEATTRVAR Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1308

Dictionary

RANGEATTRMAP Statement
Creates an attribute map that matches colors to numeric values or numeric ranges so that the colors can
be associated with the values of an input data column.
Restriction: A RANGEATTRMAP cannot be directly referenced in a legend statement.
Requirements: The RANGEATTRMAP block must appear in the global definition area of the
template between the BEGINGRAPH statement and the first LAYOUT statement. It
cannot be nested within a RANGEATTRMAP or DISCRETEATTRMAP block.
The RANGEATTRMAP block must contain at lease one RANGE statement.
The RANGEATTRVAR statement must be used to associate the range attribute map
with a data column.
Note: The RANGEATTRMAP statement defines a set of graphical properties for ranges of
data values.
See: “RANGEATTRVAR Statement” on page 1308

Syntax
RANGEATTRMAP NAME="string";
RANGE range-spec </option(s)>;
<… more-RANGE statements …>
ENDRANGEATTRMAP;

Required Argument
NAME="string "
assigns a name to the attribute definition for reference in a RANGEATTRVAR
statement.
1302 Chapter 19 • Range Attribute Map Statements

Restriction The string is case sensitive, must be enclosed in quotation marks, and
must define a name that is unique among RANGEATTRMAP names
within the template.

RANGE Statement Required Argument

range-spec
specifies a range of numeric values or a keyword, such as OTHER or MISSING.
A range of numeric values is specified in the form low-value - high-value. Both the
low value and the high value can be specified as an unformatted numeric value. A
less-than symbol (<) can be placed after the low numeric value, before the high
numeric value, or in both positions to exclude that value from the range endpoint
(similar to the VALUE statement of PROC FORMAT). If you are excluding the first
value in a range, then put the < after the low value. If you are excluding the last
value in a range, then put the < before the high value. You can also exclude both the
low and the high value.
For example, the following range does not include 0:
0 < - 100

Likewise, the following range does not include 100:

0 - < 100

If a value at the high end of one range also appears at the low end of another range
and you do not use the < exclusion notation, then the value is assigned to the first
range.
If two or more RANGE statements define colors to associate with the same numeric
values or ranges, then the first RANGE statement's settings are used. If any RANGE
statement's range overlaps another RANGE statement's range (for example, 10 - 20
and 15 - 25), then the entire attribute map is ignored and default coloring is used.
If two ranges share a common endpoint (for example, 10 - 20 and 20 - 30) and no
exclusion operator is used, then the common endpoint belongs to the lower
encountered range (10 - 20 in this case). The order of the specification does not
matter.
To set a single numeric value, specify the same value for both the low value and the
high value.
If the low value is not less than or equal to the high value, then the range
specification is invalid, and the RANGE statement is ignored in the attribute map
definition.
Note: If a range is not defined for keyword OTHER, then gaps within the attribute
map ranges are assigned the default color that is defined by the
GraphOther:ContrastColor style reference.
Rather than using a numeric value, you can specify one of the following keywords as
the low value or the high value:

MIN indicates the minimum data value for column values.

MAX indicates the highest data value for the column values.
MAXABS indicates max(abs(MIN) , abs(MAX))
NEGMAX indicates -MAX
NEGMAXABS indicates -max(abs(MIN) , abs(MAX))
RANGEATTRMAP Statement 1303

Rather than specifying a low-value-to-high-value range, you can use one of the
following keywords for the range specification:
MISSING
indicates a mapping for missing values. The visual attributes for this setting are
obtained from the GraphMissing style element. If one RANGE statement
specifies this value and another RANGE statement specifies keyword OTHER,
then the OTHER range does not include missing values.
UNDER
creates a range for all data values between the lowest mapped value and the
lowest actual data value. The visual attributes for this setting are obtained from
the GraphUnderflow style element. If one RANGE statement specifies this value
and another RANGE statement specifies keyword OTHER, then the OTHER
range does not include underflow values.
OVER
creates a range for all data between the highest mapped value and the highest
actual data value. The visual attributes for this setting are obtained from the
GraphOverflow style element. If one RANGE statement specifies this value and
another RANGE statement specifies keyword OTHER, then the OTHER range
does not include overflow values.
OTHER
creates a category for all other column values not explicitly assigned to a range.
The OTHER values can be composed of several non-contiguous ranges. The
visual attributes for this setting are obtained from the GraphOther style element.

RANGE Statement Optional Arguments

RANGEALTCOLOR=style-reference | color | GRADIENTSTEPPER(color1,color2,
num-steps, step)
specifies a single contrast color to represent the defined value range.
GRADIENTSTEPPER (color1,color2, num-steps, step)
a gradient stepper that partitions a color range into equal-sized intervals and
returns the color that is in the specified step position. The start and end colors for
the range are specified in parameters color1 and color2. The number of equal-
sized intervals is specified in parameter num-steps, and the step position for the
color to return is specified in parameter step.
Example:
rangeattrmap name="incomemap";
range min - 13000 / rangealtcolor=gradientstepper(red,green,4,1);
range 13000 < - 25000 / rangealtcolor=gradientstepper(red,green,4,2);
range 25000 < - 50000 / rangealtcolor=gradientstepper(red,green,4,3);
range 50000 < - max / rangealtcolor=gradientstepper(red,green,4,4);
endrangeattrmap;

Default The GraphDataDefault:ContrastColor style reference.

Interaction If this option is specified, then the RANGEALTCOLORMODEL=

option is ignored

RANGECOLOR=style-reference | color | GRADIENTSTEPPER(color1,color2,

num-steps, step )
specifies a single color to represent the defined value range.
1304 Chapter 19 • Range Attribute Map Statements

GRADIENTSTEPPER (color1,color2, num-steps, step )

a gradient stepper that partitions a color range into equal-sized intervals and
returns the color that is in the specified step position. The start and end colors for
the range are specified in parameters color1 and color2. The number of equal-
sized intervals is specified in parameter num-steps, and the step position for the
color to return is specified in parameter step. Example:
rangeattrmap name="incomemap";
range min - 13000 / rangecolor=gradientstepper(red,blue,4,1);
range 13000 < - 25000 / rangecolor=gradientstepper(red,blue,4,2);
range 25000 < - 50000 / rangecolor=gradientstepper(red,blue,4,3);
range 50000 < - max / rangecolor=gradientstepper(red,blue,4,4);
endrangeattrmap;

Default The GraphDataDefault:ContrastColor style reference.

Interaction If this option is specified, then the RANGECOLORMODEL= option is

ignored.

RANGEALTCOLORMODEL=style-element | (list-of-colors)
specifies either a style element or a list of one or more specific contrast colors to
represent the defined value range in this argument.
style-element
specifies the name of a style element. To display the range as a gradient ramp,
choose a style element such as TwoColorRamp, TwoColorAltRamp,
ThreeColorRamp, or ThreeColorAltRamp. The style element should contain the
following style attributes:

STARTCOLOR specifies a color for the smallest data value.

NEUTRALCOLOR specifies a color for the midpoint of the data range.
This attribute is not needed when defining a two-
color ramp.
ENDCOLOR specifies a color for the highest data value.
To display the range or single value contrast color as a color that is defined in a
style, use a style-reference (for example, GraphData1:color) to refer to a
color attribute. The following style references correspond to the keywords that
are available in this statement’s range-spec argument:

range-spec Keyword Corresponding Style Reference

MISSING GraphMissing:ContrastColor

OTHER GraphOther:ContrastColor

UNDER GraphUnderflow:ContrastColor

OVER GraphOverflow:ContrastColor

(list-of-colors)
a space-separated list of two or more color keywords that is enclosed in
parentheses.
Two colors create the endpoints of a ramp. The first color is assigned to the low
value in the range specification, and the second color is assigned to the high
RANGEATTRMAP Statement 1305

value. Three or more colors partition the range specification into n-1 equal-sized
intervals where each adjacent color pair defines a two-color ramp.

Interaction This option is ignored if the RANGEALTCOLOR= option is specified.

RANGECOLORMODEL=style-element | (list-of-colors)
specifies either a style element or a list of one or more specific colors to represent the
defined value range in this range-spec argument.
style-element
specifies the name of a style element. To display the range as a gradient ramp,
choose a style element such as TwoColorRamp, TwoColorAltRamp,
ThreeColorRamp, or ThreeColorAltRamp. The style element should contain the
following style attributes:

STARTCOLOR specifies a color for the smallest data value.

NEUTRALCOLOR specifies a color for the midpoint of the data range.
This attribute is not needed when defining a two-
color ramp.
ENDCOLOR specifies a color for the highest data value.
To display the range or single value as a color that is defined in a style, use a
style-reference (for example, GraphData1:color) to refer to a color attribute.
The following style references correspond to the keywords that are available in
this statement’s range-spec argument:

range-spec Keyword Corresponding Style Reference

MISSING GraphMissing:ContrastColor

OTHER GraphOther:ContrastColor

UNDER GraphUnderflow:ContrastColor

OVER GraphOverflow:ContrastColor

(list-of-colors)
a space-separated list of two or more color keywords that is enclosed in
parentheses.
Two colors create the endpoints of a ramp. The first color is assigned to the low
value in the range specification, and the second color is assigned to the high
value. Three or more colors partition the range specification into n–1 equal-sized
intervals, where each adjacent color pair defines a two-color ramp.

Interaction This option is ignored if the RANGECOLOR option is specified.

Details

Creating and Using a Range Attribute Map

The RANGEATTRMAP statement creates an attribute map that matches colors to
numeric values or value ranges. The attribute map can be associated with a data input
column that uses color to represent response values in a graph. Attribute maps can be
useful for controlling the application of gradient color in a graph. In addition, they
1306 Chapter 19 • Range Attribute Map Statements

enable you to map color to data values, independent of the actual data that is used in the
graph. For example, for temperature data, you can set Blue for 0 and Red for 100, even if
the values 0 and 100 are not in the data.
Defining and using the attribute map requires you to coordinate settings on several
statements:
• Use the RANGEATTRMAP statement to start the attribute definition and assign a
name to it.
• Nest within the RANGEATTRMAP block at least one RANGE statement that
specifies a numeric value or numeric range and the color to associate with that value
or range. Use one RANGE statement for each value range that you want to map. You
can use keywords like MIN and MAX in the range specification. For complete
details about the range specifications, see the RANGE statement.
• Use the RANGEATTRVAR statement to create a named association between the
defined attribute map and the input column that contains the numeric values. The
ATTRMAP= argument identifies the attribute map and the VAR= argument identifies
the input column. Use the ATTRVAR= argument to assign a name that can be used to
reference the named association in plot statements within the template.
• Reference the attribute map where needed. See “Referencing a Range Attribute Map
in Plot Statements” on page 1285.
The RANGEATTRMAP statement must be located within the BEGINGRAPH block but
outside of the outermost layout block. It cannot be nested in another RANGEATTRMAP
statement or in a RANGEATTRMAP statement. The RANGEATTRMAP block must
contain at least one RANGE statement. All values that are referenced in the attribute
map must be numeric. Data values that are not accounted for in the attribute map receive
the default color from the GraphOther, GraphOverflow, or GraphUnderflow style
element, depending on where the unassigned values are relative to the specified data
ranges.
Note: Unlike a DISCRETEATTRMAP statement, a RANGEATTRMAP statement
cannot be directly referenced in a legend. This is because the RANGE statements can
specify keywords like MIN and MAX that require a data association for
interpretation.

About the RANGE Statement

A RANGE statement within the RANGEATTRMAP block matches color to a numeric
value, or a color ramp to a numeric range. To match colors to multiple values or ranges,
specify multiple RANGE statements using the following general syntax for each of the
statements:
RANGE range-spec </option(s)>;
For a specific example, see the “Example Program” on page 1307.
When defining multiple RANGE statements, be careful not to define conflicting ranges.
A RANGE statement might be syntactically correct by itself while nevertheless
conflicting with settings in another RANGE statement.

Specifying Color in a RANGE Option

For specifying a color in one of the RANGE options, the color keywords can be any of
the following:
• a SAS color name (for example, blue)
• an RGB name (for example, CX0000FF or x0000FF)
Example: RANGEATTRMAP and RANGEATTRVAR Statements 1307

• an HLS value (for example, H14E162D)

• a gray-scale color code (for example, GRAYBB)
• an HTML color name (for example, AZURE)
• a SAS session color (for example, DMSBLUE).

Example: RANGEATTRMAP and RANGEATTRVAR

Statements

The following graph was generated by the “Example Program” on page 1307. The
example defines colors to associate with data ranges in an input column that is used in a
grouped scatter plot. The RANGEATTRMAP statement starts the attribute map
definition and assigns a name to it. The RANGE statements define the value ranges and
the colors to associate with those ranges. Because the SCATTERPLOT statement uses
the ContrastColor style attribute for the marker colors in a grouped plot, the
RANGEALTCOLOR= and RANGEALTCOLORMODEL= options are used in the
RANGE statement to define the range colors.
The highest range value specified, which is 0.002 in this example, does not have to be an
actual value in the data. The RANGEATTRVAR statement associates the attribute map
with the data column Density and assigns the name RANGEVAR to the association. The
SCATTERPLOT statement references the named association in its
MARKERCOLORGRADIENT= option.

Example Program
proc template;
define statgraph attrmap;
1308 Chapter 19 • Range Attribute Map Statements

begingraph;
entrytitle "Height and Weight Distribution" ;

/* define the attribute map and assign the name "densityrange" */

rangeattrmap name="densityrange" ;
range MIN - 0.0004 / rangealtcolor=blue ;
range 0.0004 < - 0.0014 / rangealtcolormodel=(lightpurple lightred) ;
range 0.0014 < - 0.002 / rangealtcolor=red ;
endrangeattrmap ;

/* associate the attribute map with input data column Density and
* assign variable name RANGEVAR to the named association */
rangeattrvar attrvar=rangevar var=density attrmap="densityrange" ;

/* reference the RANGEVAR attribute map in a plot statement */

layout overlay;
scatterplot x=height y=weight / markercolorgradient=rangevar
markerattrs=(symbol=squarefilled size=6px) name="scatter";
continuouslegend "scatter" / orient=vertical
halign=right title="Density";
endlayout;
endgraph;
end;

ods graphics / reset width=475px;

proc sgrender data=sashelp.gridded(where=(count>0)) template=attrmap;
run;

RANGEATTRVAR Statement
Creates a named association between a range attribute map of numeric values or value ranges and an
input data column.
Requirement: A range attribute map must be created using the RANGEATTRMAP statement.
See: “Example: RANGEATTRMAP and RANGEATTRVAR Statements” on page 1307 for
an example.

Syntax
RANGEATTRVAR ATTRVAR=attrvar-name
VAR=data-column | expression | dynamic
ATTRMAP="attrmap-name";

Required Arguments
ATTRVAR=attrvar-name
specifies a SAS name for the map. This name must be unique among the
RANGEATTRVAR statements within the template. The assigned name can be
referenced by other statements that can be associated with the range attribute map.the
attrvar-name
RANGEATTRVAR Statement 1309

VAR=data-column | expression | dynamic

specifies a numeric input data column to be associated with an attribute map at run
time. If an expression is used, then a new column of transformed values is created
and then matched with the attribute map.

Interaction If the column is not found or the column is of the wrong type for the
attribute map, then the RANGEATTRVAR statement is ignored.

ATTRMAP="attrmap-name "
specifies the name of an existing range attribute map.

Restriction The attrmap-name is case sensitive, must be enclosed in quotation

marks, and must be the name that was assigned to the attribute map in
the RANGEATTRMAP statement’s NAME= argument.

Details
The RANGEATTRVAR statement creates and names an association between colors that
are specified in a RANGEATTRMAP block and a numeric column that is in the data.
Attribute maps can be useful for controlling the application of gradient color in a graph
or specifying data values that are independent of the actual data. The name that is
assigned to the association in the RANGEATTRVAR statement is the name that plot
statements must reference to use the attribute map.
Defining and using a numeric-range attribute map requires you to coordinate settings on
several statements. For more information, see the RANGEATTRMAP statement’s
“Creating and Using a Range Attribute Map” on page 1305 .
The RANGEATTRVAR statement must be located within the BEGINGRAPH block but
outside of the outermost layout block. It cannot be nested in a RANGEATTRMAP
statement.
1310 Chapter 19 • Range Attribute Map Statements
1311

Part 12

Run-Time Programming Features

Chapter 20
Dynamics and Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1313

Chapter 21
Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1317

Chapter 22
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1321

Chapter 23
Conditional Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1333
1312
1313

Chapter 20
Dynamics and Macro Variables

Template Types on PROC TEMPLATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1313

DYNAMIC, MVAR, and NMVAR Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . 1313
Dynamics Compared to Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1315

Template Types on PROC TEMPLATE

PROC TEMPLATE supports different template types such as COLUMN, TABLE,
HEADER, FOOTER, and STATGRAPH. All of these template types support run-time
variable substitution via dynamics or macro variables. For STATGRAPH templates such
variables should be declared within the scope of the template definition before the
BEGINGRAPH block.

PROC TEMPLATE;
DEFINE STATGRAPH template-name ;

DYNAMIC variable-1<"text-1"> <... variable-n<"text-n">>;

MVAR variable-1<"text-1"> <... variable-n<"text-n">>;
NMVAR variable-1<"text-1"> <... variable-n<"text-n">>;
NOTES "text";

BEGINGRAPH;
GTL statements;
ENDGRAPH;
END;
RUN;

DYNAMIC, MVAR, and NMVAR Statements

Each of the DYNAMIC, MVAR, and NMVAR statements can define multiple variables
and an optional text-string denoting its purpose or usage. For example:

DYNAMIC YVAR "required" YLABEL "optional";

1314 Chapter 20 • Dynamics and Macro Variables

MVAR LOCATE "INSIDE or OUTSIDE" SYSDATE;

NMVAR TRANS "transparency factor";

Note: For template readability, it is helpful to adopt a naming convention for these
variables to distinguish them from actual option values or column names. Common
conventions include capitalization, or adding leading or trailing underscores to their
names.
Dynamics and macro variables can be referenced within the template definition as
• argument or option values. For example:

seriesplot x=date y=YVAR / curvelabel=YLABEL

curvelabellocation=LOCATE datatransparency=TRANS;

• parts of some text strings. For example:

entrytitle "Time Series for " YLABEL;

entryfootnote "Created on " SYSDATE;

Dynamics and run-time macro variable references cannot resolve to statement or option
keywords.
Note that macro variable references should not be prefaced with an ampersand (&) if
you want them to resolve at run time.
Macro variables defined by MVAR are strings when they resolve, as with SYMGET() in
the DATA step.
Macro variables defined by NMVAR are converted to numeric tokens when they resolve,
as with SYMGETN() in the DATA step.
The values for a dynamic variable do not have to be provided by the data source. Rather,
you can provide the values in the DYNAMIC statement in PROC SGRENDER,
specifying the values as a space delimited list, enclosed in quotation marks. Do not use
parentheses in the specification.
In the following example, the graph template specifies a dynamic variable named
TICKS, which is referenced on the XAXISOPTS= option in LAYOUT OVERLAY. The
DYNAMIC statement in PROC SGRENDER provides values for TICKS:
proc template;
define statgraph regress;
dynamic TICKS ;
begingraph;
layout overlay /xaxisopts=(linearopts=(tickvaluelist=TICKS));
scatterplot x=age y=weight;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.class template=regress;

dynamic TICKS="11 13 16" ;
run;

If your template uses a dynamic variable to specify a required attribute, such as a

variable name, and the name is misspelled or is not provided in the SGRENDER
procedure, then a warning is issued and the respective plot statement drops out of the
Dynamics Compared to Macro Variables 1315

final graph. A graph is produced, but it might be a blank graph, or it might show the
results of all statements except those that are in error.
For more information about using dynamics and macro variables in your templates, see
SAS Graph Template Language: User's Guide.

Dynamics Compared to Macro Variables

The main difference between dynamics and macro variables is how they are initialized.
For dynamics, use the DYNAMIC statement with PROC SGRENDER. For example,

proc sgrender data=sashelp.class template=timeseries;

dynamic yvar="inflation" ylabel="Inflation Rate";
run;

Values for dynamics that resolve to column names or strings should be quoted. Numeric
values should not be quoted.
For macro variables, use the current symbol table (local or global) to look up the macro
variable values at run time. For example,

%let locate=inside;
%let trans=0.3;

proc sgrender data=sashelp.class template=timeseries;

dynamic yvar="inflation" ylabel="Inflation Rate";
run;

No initialization is needed for system macro variables like SYSDATE.

1316 Chapter 20 • Dynamics and Macro Variables
1317

Chapter 21
Expressions

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1317
GTL Expressions Compared to WHERE Expressions . . . . . . . . . . . . . . . . . . . . . 1318
An Expression in Statement Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1318

Overview
In Graph Template Language (GTL), as in Base SAS, an expression is an arithmetic or
logical expression that consists of a sequence of operators, operands, and functions. An
operand is a dynamic, a macro variable, a column, a function, or a constant. An operator
is a symbol that requests a comparison, logical operation, or arithmetic calculation. In
GTL, the expression must be enclosed in an EVAL function.
Expressions can be used to set the following types of option values:
• a constant
• a column
• part of the text for ENTRYTITLE, ENTRYFOOTNOTE, and ENTRY statements
Here is an example of computing constants:

/* create reference lines at computed positions */

referenceline y=eval(mean(height)+2*std(height)) /
curvelabel="+2 STD";
referenceline y=eval(mean(height));
referenceline y=eval(mean(height)-2*std(height)) /
curvelabel="-2 STD";

Here is an example of creating a new column:

/* create a new column as a log transformation */

scatterplot x=date y=eval(log10(amount));

Here is an example of building a footnote text string:

/* create a date and time stamp as a footnote */

entryfootnote eval(put(today(),date9.)||" : "||
put(time(),timeampm8.));
1318 Chapter 21 • Expressions

When you are building a text string, you can use the || operator to concatenate the
substrings if none of the substrings are a numeric character value. If one or more of the
substrings are a numeric character value such as “1.5”, then the || operator might
produce unexpected results. In that case, use the CATS function to concatenate the
strings instead. Here is an example of using the CATS function to build a footnote text
string that includes the value of dynamic variable maxWeight, which stores a numeric
character value:
/* indicate the maximum weight in a footnote */
entryfootnote eval(cats("Maximum weight: ", maxWeight," lbs"));

GTL Expressions Compared to WHERE

Expressions
Valid GTL expressions are identical to valid WHERE expressions. See the WHERE
statement documentation in Base SAS for a comprehensive list of operators and
operands. However, GTL expressions do not perform subset operations as WHERE
expressions do. The major difference in the result of a logical GTL expression on a
column is that a Boolean value is returned for each observation without changing the
number of observations.
For example, in the following line of code, the expression for the Y= argument does not
reduce the number of observations plotted.

scatterplot x=name y=eval(height between 40 and 60);

Instead, the computed numeric column for the Y= argument consists of 0s and 1s, based
on whether each observation’s Height column value is between 40 and 60. Whenever
expressions are used to create new columns, a new column name is internally
manufactured so that it does not collide with other columns in use.

An Expression in Statement Syntax

Throughout GTL documentation, you see expression used in statement documentation:
BOXPLOT X=column | expression
Y=numeric-column | expression < /option(s)>;
For the X= argument, expression means any EVAL(expression) that results in either a
numeric or character column. An expression that yields a constant is not valid because
the X= argument does not accept constants.
Similarly, for the Y= argument, expression means any EVAL(expression) that results in a
numeric column. The expression cannot result in a character column or any constant
because the Y= argument only accepts a numeric column.
On the following REFERENCELINE statement, the X= argument can be a constant
(single line) or a column (multiple lines) that has the same data type as the axis. This
means that EVAL(expression) can result in a numeric or character column or constant
that agrees with the axis type.
REFERENCELINE X= x-axis-value | column | expression </option(s)>;
An Expression in Statement Syntax 1319

Automatic Type Conversion. Although expressions that are used in a DATA step
perform automatic type conversion, GTL expression evaluation does not. Thus, you
must use function(s) to perform required type conversions in an expression. Otherwise,
the expression generates an error condition without warning when the template is
executed.
For example, consider the following GTL expression:

if(substr(value, 1, 2) = "11")

This expression uses the SUBSTR function to determine whether the first two characters
from VALUE evaluate to the string value "11". If VALUE is a string, then the expression
works properly. However, if VALUE is numeric, then the expression generates an error
condition. For a numeric, you must convert the value to a string before passing it to the
SUBSTR function. The following modification uses the CATS function to perform the
type conversion when necessary:

if(substr(cats(value, 1, 2)) = "11")

1320 Chapter 21 • Expressions
1321

Chapter 22
Functions

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1321
SAS Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1321
SAS Functions That Can Be Used in a GTL Template . . . . . . . . . . . . . . . . . . . . 1321
SAS Functions That Can Be Used to Create Flexible Templates . . . . . . . . . . . . . 1322
Examples of Using the IFC and IFN SAS Functions . . . . . . . . . . . . . . . . . . . . . . 1322
Functions Defined Only in GTL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1324
GTL Functions Used with the EVAL Function . . . . . . . . . . . . . . . . . . . . . . . . . . 1324
Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1326
Using the TYPEOF SAS Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1326
GTL Summary Statistic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1328
Commonly Used Summary Statistic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 1328
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1329

Overview
GTL supports a large number of functions, including:
• SAS functions that can be used in the context of a WHERE expression
• functions that are defined only in GTL
• summary statistic functions

SAS Functions

SAS Functions That Can Be Used in a GTL Template

Most of the SAS functions that are available in WHERE expressions can be used in a
GTL template. These SAS functions include:
• character-handling functions
• date and time functions
• mathematical and statistical functions
1322 Chapter 22 • Functions

Not all SAS functions are available in WHERE expressions. Call routines and other
DATA-step-only functions (for example, LAG, VNAME, OPEN) are some examples of
functions that cannot be used. Not all functions that are available in WHERE
expressions are supported in GTL templates in all cases. The following form of the PUT
function is an example:
markercharacter = eval(put(amount, dollar7.2 -L))

This form results in an error when the template is compiled. However, the following
form is supported.
markercharacter = eval(put(amount, dollar7.2))

If you want to justify a string that is generated by the PUT function, use the LEFT or
RIGHT function with the PUT function as shown in the following example:
markercharacter = eval(left(put(amount, dollar7.2)))

Functions that accept null parameter values also might not be supported when you
specify a null parameter value.
For more information about SAS functions, see “Dictionary of SAS Functions and
CALL Routines” in SAS Functions and CALL Routines: Reference.

SAS Functions That Can Be Used to Create Flexible Templates

The following table shows some of the SAS functions can be used to increase the
flexibility of your template code.

SAS Function Name Description

IFC(logical-expression, "true-value", "false- Returns the character value true-value if

value" <,"missing-value">) logical-expression resolves to TRUE, false-
value if it resolves to FALSE, or missing-
value if it resolves to a missing value. The
TRUE, FALSE, and MISSING values must be
enclosed in quotation marks.

IFN(logical-expression, true-value, false- Returns the numeric value true-value if

value <,missing-value>) logical-expression resolves to TRUE, false-
value if it resolves to FALSE, or missing-
value if it resolves to a missing value.

Examples of Using the IFC and IFN SAS Functions

The IFC and IFN functions return one of two character or numeric values based on
whether a conditional expression resolves to TRUE or FALSE. They can also return an
optional third value if the conditional expression resolves to a missing value. These
functions enable you to specify a value based on a conditional expression, effectively
creating a new data column. In some cases, these functions can be used in place of IF-
THEN-ELSE statements in your template code. As with other functions, you must
enclose the IFC and IFN functions in the EVAL function.
Here is an example that uses both the IFN and IFC functions for creating a sales-based
commission chart for employees in a sales group. Each employee in the group works in
one of two sales units: Products and Services. The data for this example includes the
SAS Functions 1323

employee ID, total sales, and sales unit code for each member of the sales group. Here is
the data.
data sales;
input empID totalSales salesUnit $18;
format totalSales dollar9.;
datalines;
112876 129489.44 P
112421 169842.97 S
115331 108763.51 S
110765 181009.22 P
113722 147688.78 P
;

The TotalSales column contains the total sales for each employee. The SalesUnit column
contains a code that identifies the sales unit in which each employee works. The codes
are P for the Products unit and S for the Services unit.
Here is the output for this example.

The two bar charts show the total sales and earned commission for each employee. The
IFN function is used to compute commission for each employee based on his or her total
sales. Employees that achieved a sales total of $120,000 or more earn a commission of
5% of their total sales. All other employees earn a commission of 2.5% of their total
sales.
An axis table along the X axis shows the sales unit for each employee. The IFC function
is used to convert the P and S SalesUnit codes into more descriptive values in the axis
table. Because only two sales unit codes are used in this case, we can use the IFC
function for this purpose. This saves us from having to add a new column to the data in a
DATA step or having to create and apply a custom format to the SalesUnit column.
Here is the SAS code that defines the template and generates the graph.
proc template;
1324 Chapter 22 • Functions

define statgraph commission;

begingraph;
entrytitle "Sales-Based Commission";
layout overlay /
xaxisopts=(label="Employee ID")
yaxisopts=(label="Total Sales")
y2axisopts=(label="Commission"
linearopts=(viewmax=15000 tickvalueformat=dollar9.));
/* Generate the sales bar chart. */
barchart category=empID response=totalSales /
name="Sales" legendlabel="Total Sales" barwidth=0.3
discreteoffset=-0.2 fillattrs=graphData1;

/* Generate the commission bar chart. */

barchart category=empID
/* Use IFN to compute the commission. */
response=eval(ifn(totalSales >= 120000,
totalSales * 0.05, /* 5% if TRUE */
totalSales * 0.025)) / /* 2.5% if FALSE */
name="Commission" legendlabel="Commission"
barwidth=0.3 yaxis=y2 discreteoffset=0.2
fillattrs=graphData2;

/* Add an axis table that shows the sales unit for each
employee. */
innermargin / align=bottom;
axistable x=empID
/* Use IFC to convert the codes to meaningful
values. */
value=eval(ifc(salesUnit = 'P',"Products",
"Services")) / display=(values);
endinnermargin;
discretelegend "Sales" "Commission";
endlayout;
endgraph;
end;
run;

proc sgrender data=sales template=commission;

run;

Functions Defined Only in GTL

GTL Functions Used with the EVAL Function

The following table shows some functions that are used only in GTL. As with other
functions, these must be enclosed within an EVAL. In all these functions, column can be
either the name of a column in the input data set or a dynamic / macro variable that
resolves to such a column.
Functions Defined Only in GTL 1325

Function Name Description

COLC("string-1", "string-1"<, "string-n" …>) Converts a list of comma-separated string

values into a temporary character column.
Starting with the first maintenance release of
SAS 9.4, you can use this function to specify
values in options that accept a character
column.

COLN(n-1, n-1<, n-N…>) Converts a list of comma-separated numeric

values into a temporary numeric column. You
can use this function to specify values in
options that accept a numeric column.

COLNAME(column) Returns the case-sensitive name of the

column.

COLLABEL(column) Returns the case-sensitive label of the column.

If no label is defined for the column, then the
case-sensitive name of the column is returned.

EXISTS(item) Returns 1 if specified item exists, 0 otherwise.

If item is a column, then it tests for the
presence of the column in the input data set. If
item is a dynamic / macro variable, then it
tests whether there has been a run-time
initialization of the variable.

EXPAND(numeric-column, freq-column) Creates a new column as (numeric-column *

frequency-column) .

ASORT(column, RETAIN=ALL) Sorts all columns of the data object by the

values of column in ascending order. SORT is
an alias for ASORT.
Warning: if the RETAIN=ALL argument is
not included, column alone is sorted, not the
other columns, causing rowwise information
to be lost.
Limitation: only one sort operation (whether
an ASORT() or DSORT() function) can be
used within a single template definition.

DSORT(column, RETAIN=ALL) Sorts all columns of the data object by the

values of column in descending order.
Warning: if the RETAIN=ALL argument is
not included, column alone is sorted, not the
other columns, causing rowwise information
to be lost.
Limitation: only one sort operation (whether
an ASORT() or DSORT() function) can be
used within a single template definition.

NUMERATE(column) Returns a column that contains the ordinal

position of each observation in the input data
set (similar to an Obs column).
1326 Chapter 22 • Functions

Examples

/* arrange bars in descending order of response values */

barchartparm category=region response=eval(dsort(amount,retain=all));

/* label outliers with their position in the data set */

/* it does not matter which column is used for NUMERATE() */
boxplot x=age y=weight / datalabel=eval(numerate(age));

/* add information about the column being processed,

which is passed by a dynamic */
entrytitle "Distribution for " eval(colname(DYNVAR));

Using the TYPEOF SAS Function

The TYPEOF function returns the type of a specified column at run time.
TYPEOF(column)
This function returns the character ‘C’ if the specified column is a character column or
‘N’ if it is a numeric column.
You can use the TYPEOF function to take specific actions in your template at run time
based on the input data type. Here is an example that creates a graph of two columns and
uses the TYPEOF function to select a graph type that is appropriate for the column
types. The result returned by the TYPEOF function determines the graph type as
follows:
• If both columns are numeric, then it creates a scatter plot.
• If the X column is character and the Y column is numeric, then it creates a vertical
bar chart.
• If the X column is numeric and the Y column is character, then it swaps the category
and response columns in the BARCHART statement and orients the chart
horizontally.
Functions Defined Only in GTL 1327

Here is the output for the third case, a numeric X column and a character Y column.

Here is the SAS code.

/* Define the graph template. */
proc template;
define statgraph plot;
dynamic cat resp; /* Category and response columns. */
begingraph;
entrytitle "Graph of " eval(collabel(resp)) " and "
eval(collabel(cat));
layout overlay;
/* If cat and resp are numeric, then generate a scatter plot.
Otherwise, generate a bar chart. */
if (typeof(cat) = "N" and typeof(resp) = "N")
scatterplot x=cat y=resp;
else
/* If cat is a character column, then generate a vertical bar
chart. Otherwise, generate a horizontal bar chart. */
if (typeof(cat) = "C")
barchart category=cat response=resp / stat=mean;
else
barchart category=resp response=cat /
stat=mean orient=horizontal;
endif;
endif;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=plot;

dynamic cat="MPG_CITY" resp="TYPE";
run;

Note: See “Functions Defined Only in GTL” on page 1324 for information about the
COLLABEL function.
1328 Chapter 22 • Functions

GTL Summary Statistic Functions

Commonly Used Summary Statistic Functions

The following functions return a numeric constant, based on a summary operation on a
numeric column. The results are the same as if the corresponding statistics were
requested with PROC SUMMARY. These functions take a single argument that resolves
to the name of a numeric column. These functions take precedence over similar multi-
argument DATA step functions.

number = EVAL(function-name(numeric-column))

Function Name Description

CSS Corrected sum of squares

CV Coefficient of variation

KURTOSIS Kurtosis

LCLM One-sided confidence limit below the mean

MAX Largest (maximum) value

MEAN Mean

MEDIAN Median (50th percentile)

MIN Smallest (minimum) value

N Number of nonmissing values

NMISS Number of missing values

P1 1st percentile

P5 5th percentile

P25 25th percentile

P50 50th percentile

P75 75th percentile

P90 90th percentile

P95 95th percentile

P99 99th percentile

GTL Summary Statistic Functions 1329

number = EVAL(function-name(numeric-column))

Function Name Description

PROBT p-value for Student’s t statistic

Q1 First quartile

Q3 Third quartile

QRANGE Interquartile range

RANGE Range

SKEWNESS Skewness

STDDEV Standard deviation

STDERR Standard error of the mean

SUM Sum

SUMWGT Sum of weights

T Student’s t statistic

UCLM One-sided confidence limit above the mean

USS Uncorrected sum of squares

VAR Variance

Example
The following example uses GTL summary statistic functions to dynamically construct
reference lines and a table of statistics for a numeric variable, which is supplied at run
time.
1330 Chapter 22 • Functions

Here is the graph for this example.

Here is the SAS code.

proc template;
define statgraph expression;
dynamic NUMVAR "required";
begingraph;
entrytitle "Distribution of " eval(colname(NUMVAR));
layout overlay / xaxisopts=(display=(ticks tickvalues line));
histogram NUMVAR;

/* create reference lines at computed positions */

referenceline x=eval(mean(NUMVAR)+2*std(NUMVAR)) /
lineattrs=(pattern=dash) curvelabel="+2 STD";
referenceline x=eval(mean(NUMVAR)) /
lineattrs=(thickness=2px) curvelabel="Mean";
referenceline x=eval(mean(NUMVAR)-2*std(NUMVAR)) /
lineattrs=(pattern=dash) curvelabel="-2 STD";

/* create inset */
layout gridded / columns=2 order=rowmajor
autoalign=(topleft topright) border=true;
entry halign=left "N";
entry halign=left eval(strip(put(n(NUMVAR),12.0)));
entry halign=left "Mean";
entry halign=left eval(strip(put(mean(NUMVAR),12.2)));
entry halign=left "Std Dev";
entry halign=left eval(strip(put(stddev(NUMVAR),12.2)));
endlayout;
endlayout;
endgraph;
end;
run;
GTL Summary Statistic Functions 1331

proc sgrender data=sashelp.heart template=expression;

dynamic numvar="MRW";
run;
1332 Chapter 22 • Functions
1333

Chapter 23
Conditional Logic

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1333
Conditional Logic Determines Statement Rendering . . . . . . . . . . . . . . . . . . . . . . 1334
GTL Does Not Provide ELSE IF Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1335

Overview
GTL supports conditional logic that enables you to include or exclude one or more GTL
statements at run time:
IF (condition)
GTL-statement(s);
ELSE
GTL-statement(s);
ENDIF;
The IF statement requires an ENDIF statement. The IF block can be placed anywhere
within the BEGINGRAPH / ENDGRAPH block.
The condition is an expression that evaluates to a numeric constant, where all numeric
constants other than 0 and MISSING are true. There is an implied EVAL(condition), so
it is not necessary to include an EVAL as part of the condition.
Examples:
/* test a computed value */
if (weekday(today()) in (1 7))
entrytitle "Run during the work week";
else
entrytitle "Run during the weekend";
endif;

/* test for the value a numeric dynamic */

if (ADDREF > 0)
referenceline y=1;
referenceline y=0;
referenceline y=-1;
endif;
1334 Chapter 23 • Conditional Logic

/* test for the value a character dynamic */

if (upcase(ADDREF) =: "Y")
referenceline y=1;
referenceline y=0;
referenceline y=-1;
endif;

/* test whether a dynamic is initialized */

if (exists(ADDREF))
referenceline y=1;
referenceline y=0;
referenceline y=-1;
endif;

Conditional Logic Determines Statement

Rendering
The GTL conditional logic is used only to determine which statements are rendered, not
to control what is in the data object. In the following example, the data object contains
columns for Date, Amount, and LOG(AMOUNT), but only one scatter plot is created.

if (LOGFLAG)
scatterplot x=date y=amount;
else
scatterplot x=date y=log(amount);
endif;

Also, it is seldom necessary to test for the existence of option values set by columns or
dynamics. Consider the following statement:

scatterplot x=date y=amount / group=GROUPVAR;

This SCATTERPLOT statement is equivalent to the following code because option

values that are set by columns that do not exist or dynamics that are uninitialized simply
“drop out” at run time and do not produce errors or warnings:

if (exists(GROUPVAR))
scatterplot x=date y=amount / group=GROUPVAR;
else
scatterplot x=date y=amount;
endif;

The GTL code that is conditional must be complete statements, or complete blocks of
statements, or both. The following IF block produces a compile error because there are
more LAYOUT statements than ENDLAYOUT statements:

/* this IF block produces a compile error */

if (exists(SQUAREPLOT))
layout overlayequated / equatetype=square;
else
layout overlay;
endif;
GTL Does Not Provide ELSE IF Syntax 1335

scatterplot x=XVAR y=YVAR;

endlayout;

This is the correct conditional construct:

if (exists(SQUAREPLOT))
layout overlayequated / equatetype=square;
scatterplot x=XVAR y=YVAR;
endlayout;
else
layout overlay;
scatterplot x=XVAR y=YVAR;
endlayout;
endif;

GTL Does Not Provide ELSE IF Syntax

The GTL does not provide ELSE IF syntax, but you can create a nested IF/ELSE block
as follows:
IF (condition-1)
GTL-statement(s);
ELSE
IF (condition-2)
GTL-statement(s);
ELSE
GTL-statement(s);
ENDIF;
ENDIF;
1336 Chapter 23 • Conditional Logic
1337

Part 13

Appendixes

Appendix 1
Syntax Conventions and Argument Value Types . . . . . . . . . . . . . . . . 1339

Appendix 2
Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . 1343

Appendix 3
Display Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347

Appendix 4
SAS Formats Not Supported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353

Appendix 5
Generalized Macro for BOXPLOTPARM Data . . . . . . . . . . . . . . . . . . . 1357

Appendix 6
Memory Management for ODS Graphics . . . . . . . . . . . . . . . . . . . . . . . . 1363
1338
1339

Appendix 1
Syntax Conventions and
Argument Value Types

Syntax Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1339

Value Types for Statement Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1339

Syntax Conventions
< > ==> optional
a | b ==> either a or b

Value Types for Statement Options

The default value of an option depends on the template definition that is in use, and the
implementation of that option depends on the ODS destination that formats the output.
In addition, if you are creating HTML output, then the implementation of an attribute
depends on the browser that you use.
This section describes the value types that are available for GTL statement options.
boolean
specifies a literal value that resolves to true or false. The following table lists literal
values that resolve to true or false.

Table A1.1 Boolean Values

Values That Resolve To True Values That Resolve To False

ON OFF

_ON_ _OFF_

TRUE FALSE

YES NO
1340 Appendix 1 • Syntax Conventions and Argument Value Types

Values That Resolve To True Values That Resolve To False

_YES_ _NO_

1 0

color
specifies a string that identifies a color. A color can be one of the following:
• any of the color names that are supported by SAS. See “Color-Naming Schemes”
in SAS Graph Template Language: User's Guide.
• one of the colors that exists in the SAS session when the style template is used,
such as DMSBLACK or DMSCYAN. (Use these color specifications only if you
are running SAS in the windowing environment.)
• an English description of an Hue/Light/Saturation (HLS) value. Such
descriptions use a combination of words to describe the lightness, the saturation,
and the hue (in that order). You can use the Color Naming System to form a color
by doing one of the following:
• combining a chromatic hue with a lightness, a saturation, or both
• combining the achromatic hue gray with a lightness
• combining the achromatic hue black or white without qualifiers.
• combining words to form a wide variety of colors, such as light vivid green,
dark vivid orange, or light yellow.
• specify hues that are intermediate between two neighboring colors. To do so,
combine one of the following adjectives with one of its neighboring colors:
brownish, greenish, purplish, or yellowish (for example, bluish purple or reddish
orange).
column
specifies a column variable that contains either double-precision values or string
values, or a dynamic variable that refers to such a column.
See also: integer-column, numeric-column, and string-column.
dimension
specifies a nonnegative number. The number can be followed by one of the
following optional units of measure:

Table A1.2 Units for Dimension

Unit Description

CM centimeters

IN inches

MM millimeters

PCT or % percentage

PT point size (72 points = 1 inch)

Value Types for Statement Options 1341

Unit Description

PX pixels

expression
specifies a selective, relational, or logical program structure that calculates values
when those values are not stored in the data. The expression must be specified as an
EVAL() argument. The following shows the structure of an EVAL() argument:

x = EVAL(expression)

The expression returns a number and can be formed with consonants, data columns,
dynamic variables, functions, or other expressions. The following example uses the
data column Time and the SGE functions MEAN and ACF:

EVAL(MEAN(Time) + ACF(Time, NLags=10))

For more information about expressions, see Chapter 21, “Expressions,” on page
1317.
format
specifies a SAS format or a user-defined format.

integer, integer-column
specifies a member of the set of positive whole numbers, negative whole numbers,
and zero.
An integer column specifies a column that contains integer values, or a dynamic
variable that refers to such a column.
line-pattern-name, line-pattern-number
specifies a string value of a line pattern, a numeric value of a line pattern, a dynamic
variable that contains such a string or number, or a style reference to a line pattern.
Line patterns are chosen for discriminability. Because of different densities, equal
weighting is impossible for lines of the same thickness. Instead, line patterns are
ordered to provide a continuum of weights, which is useful when displaying
confidence bands.
For details about line attributes, see “Line Options” on page 1349.
marker-name
specifies a string value of a maker symbol, a dynamic variable that contains a marker
symbol, or a style reference to a marker symbol.
For details about marker attributes, see “Marker Options” on page 1350.
number, numeric-column
specifies a value, a dynamic variable that contains a double-precision value, an
expression that resolves to a double-precision value, or a style reference to a double-
precision value.
A numeric-column specifies a column that contains double-precision values, or a
dynamic variable that refers to such a column.
string, string-column
specifies a quoted character string.
A string-column specifies a column that contains string values, or a dynamic variable
that refers to such a column.
1342 Appendix 1 • Syntax Conventions and Argument Value Types

Note: For quoted character string options in the GTL, a space enclosed in quotation
marks (" " or ' ') and empty quotation marks ("" or '') are not equivalent. A space
enclosed in quotation marks specifies a blank space or a missing string value.
Empty quotation marks has the same effect as not setting the option. To specify a
blank space or missing value in a quoted string option, use a space enclosed in
quotation marks (" " or ' ').
style-reference
specifies a reference to an attribute that is defined in a style element.
In the ODS Graphics templates that SAS provides, options for plot features are
specified with a style reference in the form style-element:attribute, rather than a
specific value. For example, the symbol, color, and size of markers for a basic scatter
plot is specified in a SCATTERPLOT statement as follows:

scatterplot x=X y=Y /

markersymbol=GraphDataDefault:markersymbol
markercolor=GraphDataDefault:contrastcolor
markersize=GraphDataDefault:markersize

The above style references guarantee a common appearance for markers used in all
basic scatter plots. For non-grouped data, the marker appearance is controlled by the
GraphDataDefault style element in the style template that you specify.
In order to create your own style template, or to modify a style template to use with
ODS Graphics, you need to understand the relationship between style elements and
graph features. For more information, see the usage guide.
1343

Appendix 2
Reserved Keywords and Unicode
Values

Overview
The tables in this section show some of the reserved keywords and Unicode values that
can be used with the UNICODE text command. For information about rendering
Unicode characters, see “Managing the String on Text Statements” in SAS Graph
Template Language: User's Guide.
Note the following:
• Keywords and Unicode values are not case-sensitive: "03B1"x is the same code point
as "03b1"x.
• The word blank is the keyword for a blank space.

Lowercase Greek Letters

Keyword Glyph Unicode Description

alpha α 03B1 lowercase alpha

beta β 03B2 lowercase beta

gamma γ 03B3 lowercase gamma

delta δ 03B4 lowercase delta

epsilon ε 03B5 lowercase epsilon

zeta ζ 03B6 lowercase zeta

eta η 03B7 lowercase eta

theta θ 03B8 lowercase theta

iota ι 03B9 lowercase iota

kappa κ 03BA lowercase kappa

lambda λ 03BB lowercase lambda

1344 Appendix 2 • Reserved Keywords and Unicode Values

Keyword Glyph Unicode Description

mu μ 03BC lowercase mu

nu ν 03BD lowercase nu

xi ξ 03BE lowercase xi

omicron ο 03BF lowercase omicron

pi π 03C0 lowercase pi

rho ρ 03C1 lowercase rho

sigma σ 03C3 lowercase sigma

tau τ 03C4 lowercase tau

upsilon υ 03C5 lowercase upsilon

phi φ 03C6 lowercase phi

chi χ 03C7 lowercase chi

psi ψ 03C8 lowercase psi

omega ω 03C9 lowercase omega

Uppercase Greek Letters

Table A2.1 Uppercase Greek Letters

Keyword Glyph Unicode Description

alpha_u Α 0391 uppercase alpha

beta_u Β 0392 uppercase beta

gamma_u Γ 0393 uppercase gamma

delta_u Δ 0394 uppercase delta

epsilon_u Ε 0395 uppercase epsilon

zeta_u Ζ 0396 uppercase zeta

eta_u Η 0397 uppercase eta

theta_u Θ 0398 uppercase theta

Reserved Keywords and Unicode Values 1345

Keyword Glyph Unicode Description

iota_u Ι 0399 uppercase iota

kappa_u Κ 039A uppercase kappa

lambda_u Λ 039B uppercase lambda

mu_u Μ 039C uppercase mu

nu_u Ν 039D uppercase nu

xi_u Ξ 039E uppercase xi

omicron_u Ο 039F uppercase omicron

pi_u Π 03A0 uppercase pi

rho_u Ρ 03A1 uppercase rho

sigma_u Σ 03A3 uppercase sigma

tau_u Τ 03A4 uppercase theta

upsilon_u Υ 03A5 uppercase upsilon

phi_u Φ 03A6 uppercase phi

chi_u Χ 03A7 uppercase chi

psi_u Ψ 03A8 uppercase psi

omega_u Ω 03A9 uppercase omega

Special Characters

Keyword Glyph Unicode Description

prime ´ 00B4 single prime sign

bar ̅ 0305 combining overline*

bar2 ̿ 033F combining double

overline*

tilde ̃ 0303 combining tilde*

1346 Appendix 2 • Reserved Keywords and Unicode Values

Keyword Glyph Unicode Description

hat ̂ 0302 combining

circumflex accent*

* This is an overstriking character that requires a Unicode font to render properly.

1347

Appendix 3
Display Attributes

General Syntax for Attribute Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347

Attributes Available for the Attribute Options . . . . . . . . . . . . . . . . . . . . . . . . . . . 1348
Fill Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1348
Line Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1349
Marker Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1350
Text Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1351
Available Line Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1352

General Syntax for Attribute Options

Most statements provide options that enable you to specify attributes for the fills, lines,
data markers, or text that is used in the display. For example, many plots provide a
DATALABELATTRS= option that specifies the attributes of the data labels. This
appendix discusses the general syntax for those options and the valid values for they
accept.
A statement’s attribute options use the following general syntax:
ATTRSOPTIONNAME= style-element | style-element (options) | (options)
style-element
Name of a style element. Only style attributes relevant for rendering the fill, line,
data marker, or text are used.

Example
ATTRSOPTIONNAME=GRAPHVALUETEXT

style-element (options)
Name of a style element, plus individual options to be used as style overrides. Any
options not specified are derived from the specified style element.

Example
ATTRSOPTIONNAME=GRAPHVALUETEXT(SIZE=10pt)

(options)
Individual options. Any options not specified are derived from the default style
element.
1348 Appendix 3 • Display Attributes

Example
ATTRSOPTIONNAME=(FAMILY="Arial" SIZE=10pt)
ATTRSOPTIONNAME=(FAMILY=GraphValueText:FontFamily)

Depending on the attribute option used, the options might be fill options , line options ,
marker options , or text options .
In general, any relevant attribute that is not specified defaults to an internal value, which
is typically derived from the style element that you specify for the attributes. When
choosing a style element, you should use an element of the correct type. See “Graph
Style Elements Used by ODS Graphics” in SAS Graph Template Language: User's
Guide for a list of style elements and their types.

Attributes Available for the Attribute Options

Depending on the attribute option used on a statement, the available attributes might be
fill options , line options , marker options , or text options .

Fill Options
When specifying the attributes for an area fill, the fill options can be one or more of the
following settings. The option must be enclosed in parentheses and specified as a
name=value pair. The value can be a style reference in the form style-element:style-
attribute.
COLOR=style-reference | color
specifies the fill color . If you use a style reference, then the style attribute should be
a valid attribute such as COLOR, CONTRASTCOLOR, STARTCOLOR,
NEUTRAL, ENDCOLOR. The convention is to use the COLOR attribute for fill
areas.
If you use a color, then SAS accepts color names, such as RED, or color codes, such
as CXFF0000 or #FF0000. Color names must not exceed 64 characters. Color codes
must not exceed 8 characters and must be in a valid SAS color-naming scheme, such
as RGB, CMYK, HLS, or HSV (HSB).
TRANSPARENCY=number
specifies the degree of the transparency of the filled area. This setting enables you to
set the transparency for the filled elements of some graph types. You can set just this
fill transparency, or set the fill independently of the other transparent elements in the
graph. For example, you can use this setting to set the transparency level for the
filled bars of a bar chart, and use the bar chart’s DATATRANSPARENCY= option to
set a different transparency level for the bar outlines.

Default The same as the setting of the statement’s DATATRANSPARENCY=

option.

Range 0–1, where 0 is opaque and 1 is entirely transparent

Interaction This setting overrides the statement’s DATATRANSPARENCY=

setting for the fills but not for the outlines.

Example fillattrs=(transparency=0.5)
Attributes Available for the Attribute Options 1349

Line Options
When specifying the attributes for a line, the available line options can be any one or
more of the following settings. The options must be enclosed in parentheses, and each
option is specified as a name=value pair. In all cases, the value can be a style reference
in the form style-element:style-attribute.
COLOR=style-reference | color
specifies the line color. If you use a style reference, then the style attribute should be
a valid attribute such as COLOR, CONTRASTCOLOR, STARTCOLOR,
NEUTRAL, ENDCOLOR. The convention is to use CONTRASTCOLOR for lines.
If you specify a style element that does not have a CONTRASTCOLOR attribute,
then the element’s COLOR attribute is used.
If you use a color, then SAS accepts color names, such as RED, or color codes, such
as CXFF0000 or #FF0000. Color names must not exceed 64 characters. Color codes
must not exceed 8 characters and must be in a valid SAS color-naming scheme, such
as RGB, CMYK, HLS, or HSV (HSB).
PATTERN=style-reference | line-pattern-name | line-pattern-number
specifies the line pattern. If you use a style reference, then the style attribute should
be LINESTYLE.
Line patterns can be specified as a pattern name or pattern number.

Valid pattern numbers range from 1 to 46. Not all pattern numbers have names. See
“Available Line Patterns” on page 1352 for a list of all possible line patterns. We
recommend that you use the named patterns because they have been optimized to
provide good discriminability when used in the same plot.

Note Anti-aliasing might alter the appearance of some line patterns that have fine
detail such as line patterns 33 through 46. For example, if you specify the
color black and the pattern 33 for a line, and anti-aliasing is enabled, then the
line might appear gray. In that case, you can use the following command to
disable anti-aliasing in order to show the line detail:
ods graphics / antialias=off;

THICKNESS=style-reference | dimension
specifies the line thickness. If you use a style reference, then the style attribute
should be LINETHICKNESS.
1350 Appendix 3 • Display Attributes

See “dimension” on page 1340

Marker Options
When you specify the attributes for a data marker, the available marker options can be
any one or more of the following settings. You must enclose the options in parentheses,
and you specify each option as a name=value pair. In all cases, the value can be a style
reference in the form style-element:style-attribute.
COLOR=style-reference | color
specifies the color of the marker. If you use a style reference, then the style attribute
should be a valid attribute such as COLOR, CONTRASTCOLOR, STARTCOLOR,
NEUTRAL, or ENDCOLOR. The convention is to use CONTRASTCOLOR for
markers. For grouped data, this option keeps all markers the same color and the
marker symbol alone distinguishes the group values.
If you use a color, then SAS accepts color names, such as RED, or color codes, such
as CXFF0000 or #FF0000. Color names must not exceed 64 characters. Color codes
must not exceed 8 characters and must be in a valid SAS color-naming scheme, such
as RGB, CMYK, HLS, or HSV (HSB).

Restriction This option has no effect on marker symbols that are created with the
SYMBOLIMAGE statement.

See “SYMBOLIMAGE Statement” on page 1180

SIZE=style-reference | dimension
specifies the marker size (both width and height). If you use a style reference, then
the style attribute should be MARKERSIZE.

See “dimension” on page 1340

SYMBOL=style-reference | marker-name
specifies the name of the marker. If you use a style reference, then the style attribute
should be MARKERSYMBOL. The following SAS symbols are supported:

You can also specify the names of symbols that are created by the SYMBOLCHAR
and SYMBOLIMAGE statements.

See “SYMBOLCHAR Statement” on page 1173

“SYMBOLIMAGE Statement” on page 1180

Attributes Available for the Attribute Options 1351

TRANSPARENCY=number
specifies the degree of transparency for the plot markers.

Default The transparency that is specified by the DATATRANSPARENCY=

option, which is 0 by default.

Range 0–1, where 0 is opaque and 1 is entirely transparent

Interaction This suboption overrides the DATATRANSPARENCY= option for the

plot markers only.

WEIGHT=NORMAL | BOLD
specifies the marker weight.

Restriction This option has no effect on marker symbols that are created with the
SYMBOLCHAR and SYMBOLIMAGE statements.

See “SYMBOLCHAR Statement” on page 1173

“SYMBOLIMAGE Statement” on page 1180

Text Options
When specifying the attributes for text, the available text options can be any one or more
of the following settings. The options must be enclosed in parentheses, and each option
is specified as a name=value pair. In all cases, the value can be a style reference in the
form style-element:style-attribute.
COLOR=style-reference | color
specifies the color of the text. If you use a style reference, then the style attribute
should be a valid attribute such as COLOR, CONTRASTCOLOR, STARTCOLOR,
NEUTRAL, ENDCOLOR. The convention is to use COLOR for text.
If you use a color, then SAS accepts color names, such as RED, or color codes, such
as CXFF0000 or #FF0000. Color names must not exceed 64 characters. Color codes
must not exceed 8 characters and must be in a valid SAS color-naming scheme, such
as RGB, CMYK, HLS, or HSV (HSB).
FAMILY=style-reference | "string"
specifies the font family of the text. If you use a style reference, then the style
attribute should be FONTFAMILY.
SIZE=style-reference | dimension
specifies the font size of the text. If you use a style reference, then the style attribute
should be FONTSIZE.

Restriction The font size cannot be less than the minimum font size in SAS, which
is determined by the SAS system. If you specify a font size that is less
than the minimum font size, the minimum size is used instead.

See “dimension” on page 1340

STYLE=style-reference | NORMAL | ITALIC

specifies the font style of the text. If you use a style reference, then the style attribute
should be FONTSTYLE.
1352 Appendix 3 • Display Attributes

WEIGHT=style-reference | NORMAL | BOLD

specifies the font weight of the text. If you use a style reference, then the style
attribute should be FONTWEIGHT.

Available Line Patterns

The following line patterns can be used with the Graphics Template Language. A line
pattern can be specified by its number or name. Not all patterns have names. We
recommend that you use the named patterns because they have been optimized to
provide good discriminability when used in the same plot.
1353

Appendix 4
SAS Formats Not Supported

Using SAS Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353

Unsupported Numeric Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353
Unsupported Date and Time Formats Related to ISO 8601 . . . . . . . . . . . . . . . . . 1354
Other Unsupported Date and Time Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1354
Unsupported Currency Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1355
Unsupported User-Defined Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1355

Using SAS Formats

SAS formats can be assigned to input data columns with the FORMAT statement of the
SGRENDER procedure. Also, several GTL statement options enable a SAS format as an
option value. Examples include the TICKVALUEFORMAT= option for formatting axis
tick values, and the TIPFORMAT= option for formatting data tips.
Not all SAS formats are supported in the GTL or with the SGPLOT, SGSCATTER,
SGPANEL, and SGRENDER procedures. The tables in the following sections show the
character and numeric SAS formats that are not supported.
When the GTL encounters an unsupported format, a note similar to the following is
written to the SAS log:

TICKVALUEFORMAT=bestx. is invalid. The format is invalid or unsupported. The

default will be used.

Unsupported Numeric Formats

The following numeric formats are not supported in the GTL:

BESTD BESTX D FLOAT FRACT

FREE IB IBR IEEE IEEER

1354 Appendix 4 • SAS Formats Not Supported

ODDSR PCPIB PD PIB PIBR

PK RB SSN WORDF WORDS

Z ZD

Unsupported Date and Time Formats Related to

ISO 8601
The following date and time formats are not supported in the GTL:

$N8601B $N8601BA $N8601E $N8601EA $N8601EH

$N8601EX $N8601H $N8601X B8601DA B8601DN

B8601DT B8601DZ B8601LZ B8601TM B8601TZ

E8601DA E8601DN E8601DT E8601DZ E8601LZ

E8601TM E8601TZ IS8601DA IS8601DN IS8601DT

IS8601DZ IS8601LZ IS8601TM IS8601TZ

Other Unsupported Date and Time Formats

The following date and time formats are not supported in the GTL:

HDATE HEBDATE JDATEMDW JDATEMNW JDATEWK

JDATEYDW JDATEYM JDATEYMD JDATEYMW JDATEYT

JDATEYTW JNENGO JNENGOT JNENGOTW JNENGOW

JTIMEH JTIMEHM JTIMEHMS JTIMEHW JTIMEMW

JTIMESW MDYAMPM MINGUO NENGO NLDATEYQ

NLDATEYR NLDATEYW NLDATMYQ NLDATMYR NLDATMYW

NLSTRMON NLSTRQTR NLSTRWK PDJULG PDJULI

TWMDY XYYMMDD YYQZ

Unsupported User-Defined Formats 1355

Unsupported Currency Formats

The following currency formats are not supported in the GTL:

EURFRATS EURFRBEF EURFRCHF EURFRCZK EURFRDEM

EURFRDKK EURFRESP EURFRFIM EURFRFRF EURFRGBP

EURFRGRD EURFRHUF EURFRIEP EURFRITL EURFRLUF

EURFRNLG EURFRNOK EURFRPLZ EURFRPTE EURFRROL

EURFRRUR EURFRSEK EURFRSIT EURFRTRL EURFRYUD

EURTOATS EURTOBEF EURTOCHF EURTOCZK EURTODEM

EURTODKK EURTOESP EURTOFIM EURTOFRF EURTOGBP

EURTOGRD EURTOHUF EURTOIEP EURTOITL EURTOLUF

EURTONLG EURTONOK EURTOPLZ EURTOPTE EURTOROL

EURTORUR EURTOSEK EURTOSIT EURTOTRL EURTOYUD

Unsupported User-Defined Formats

In the second maintenance release of SAS 9.4 and in earlier releases, ODS Graphics
does not support Unicode values in user-defined formats. Starting with the third
maintenance release of SAS 9.4, ODS Graphics supports Unicode values in user-defined
formats only if they are preceded by the (*ESC*) escape sequence as shown in the
following example.
"(*ESC*){unicode beta}"

ODS Graphics does not support the use of a user-defined ODS escape character to
escape Unicode values in user-defined formats.
For an example of how to use Unicode values in user-defined formats with ODS
Graphics, see “Formatting the Tick Values on a Discrete Axis” in SAS Graph Template
Language: User's Guide.
1356 Appendix 4 • SAS Formats Not Supported
1357

Appendix 5
Generalized Macro for
BOXPLOTPARM Data

The following SAS code is a generalized macro for computing input data for
BOXPLOTPARM.
%macro boxcompute(indsn=,x=,y=,outdsn=boxdata,datalabel=,
qntldef=5,table=no);
/* NOTE: INDSN, X and Y are required parameters, where
INDSN = input SAS data set
X = categorical variable (num or char)
Y = response variable (num)
OUTDSN = output dataset. It contains these variables:
STAT: Statistic names for BOXPLOTPARM
VALUE: values for STAT type
X: X variable values
DATALABEL: outlier labels from the DATALABEL= variable
N, Mean, Median, Std if TABLE=YES
DATALABEL= variable used to label outliers (num or char)
QNTLDEF = 1|2|3|4|5
(how to compute quantiles - see PROC SUMMARY)
TABLE = YES | NO
(add additional data to build table of statistics)
*/
%macro varinfo(dsid,varname,role,rc);
/* utility macro for obtaining variable info */
%local varnum;
%if %length(&varname)=0 %then %do;
%let &rc=0; %return;
%end;
%let varnum=%sysfunc(varnum(&dsid,&varname));
%if &varnum > 0 %then %do;
%let &role.label=%sysfunc(varlabel(&dsid,&varnum));
%if %length(&&&role.label)=0 %then
%let &role.label=%sysfunc(varname(&dsid,&varnum));
%let &role.fmt=%sysfunc(varfmt(&dsid,&varnum));
%let &rc=0;
%end;
%else %do;
%put ERROR: %upcase(&role) variable &varname not found.;
%let &rc=1;
%end;
%mend varinfo;

/* validate dataset and variables */

%local dsid ylabel xlabel datalabellabel
yfmt xfmt datalabelfmt rc_y rc_x rc_d;
1358 Appendix 5 • Generalized Macro for BOXPLOTPARM Data

%let dsid=%sysfunc(open(&indsn));
%if &dsid %then %do;
%varinfo(&dsid,&y,Y,rc_y)
%varinfo(&dsid,&x,X,rc_x)
%if %length(&datalabel) %then
%varinfo(&dsid,&datalabel,DATALABEL,rc_d);
%else %let rc_d=0;
%let dsid=%sysfunc(close(&dsid));
%if &rc_y or &rc_x or &rc_d %then %return;
%end;
%else %do;
%put ERROR: Input dataset &indsn not found.;
%return;
%end;

/* compute basic summary statistics */

proc summary data=&indsn(rename=(&y=VALUE &x=X))
nway qntldef=&qntldef;
class x;
var value;
output out=summary(drop=_type_ _freq_) n=N mean=Mean
median=Median q1=Q1 q3=Q3 std=STD / noinherit;
run;
proc sort data=&indsn(keep=&x &y &datalabel)
%if %length(&datalabel) %then
out=sorted(rename=(&x=X &y=VALUE &datalabel=DATALABEL));
%else out=sorted(rename=(&x=X &y=VALUE));
; by &x;
run;
/* compute fences, MIN, MAX and any outliers for X values */
data outliers;
length STAT $10;
%if %length(&datalabel) %then
%do;
keep STAT X VALUE DATALABEL;
label VALUE="&ylabel" X="&xlabel"
DATALABEL="&datalabellabel";
format VALUE &yfmt X &xfmt DATALABEL &datalabelfmt;
%end;
%else
%do;
keep STAT X VALUE;
label VALUE="&ylabel" X="&xlabel";
format VALUE &yfmt X &xfmt;
%end;
retain lowerFence upperFence farLowerFence farUpperFence
tempmin tempmax;
merge sorted summary; by x;

/* perform computations for each X value */

if first.X then do;
lowerFence=q1-((q3-q1)*1.5);
upperFence=q3+((q3-q1)*1.5);
farLowerFence=q1-((q3-q1)*3);
farUpperFence=q3+((q3-q1)*3);
/* these computations for MIN and MAX result
Generalized Macro for BOXPLOTPARM Data 1359

in the same values produced by the BOXPLOT

statement, however they can be modified to
satify other statistical definitions */
if value <= upperFence then tempmax=value;
else tempmax=.;
if value => lowerFence then tempmin=value;
else tempmin=.;
end;
/* recompute MAX and MIN for each obs */
if 0 <= sum(upperFence,-value) then
tempmax=max(tempmax,value);
if 0 <= sum(value,-lowerFence) then
tempmin=min(tempmin,value);
/* write out both types of outliers */
if value < farLowerFence or value > farUpperFence then do;
stat="FAROUTLIER"; output;
end;
else if value < LowerFence or value > UpperFence then do;
stat="OUTLIER"; output;
end;
/* write out MIN and MAX for each X value */
if last.X then do;
value=tempmin; stat="MIN"; output;
value=tempmax; stat="MAX"; output;
end;
run;
/* transpose the stats into the columns
required by BOXPLOTPARM */
data transpose(keep=x stat value);
length STAT $10;
set summary;
array stats{*} n--std;
do i=1 to dim(stats);
stat=upcase(vname(stats{i}));
VALUE=stats{i};
output;
end;
run;
/* interleave the obs by the X variable */
data &outdsn;
set transpose outliers; by X;
run;
/* merge the output stats for building a stat table */
%if %upcase(&table)=YES %then %do;
data &outdsn;
merge &outdsn summary; by X;
run;
%end;
%mend boxcompute;

Here is the macro invocation to produce the data for the graph shown in the section
“Example: BOXPLOTPARM Statement” on page 366 .
%boxcompute(indsn=sashelp.cars,x=type,y=mpg_city,
datalabel=make);

proc template;
1360 Appendix 5 • Generalized Macro for BOXPLOTPARM Data

define statgraph boxplotparm1;

begingraph;
entrytitle "City Mileage for Vehicle Types";
layout overlay;
boxplotparm y=value x=x stat=stat /
datalabel=datalabel spread=true ;
endlayout;
endgraph;
end;
run;

proc sgrender data=boxdata template=boxplotparm1;

run;

The following figure and code show an example of a “table” of statistics with
BLOCKPLOT statements.

%boxcompute(indsn=sashelp.cars,x=type,y=mpg_highway,
outdsn=boxdata2,table=yes);

proc template;
define statgraph boxplotparm2;
begingraph;
entrytitle "Highway Mileage for Vehicle Types";
layout overlay / xaxisOpts=(offsetMin=0.08 offsetMax=0.08);
innerMargin / align=top;
blockplot x=x block=n /
display=(values label outline) valuehalign=center
labelattrs=graphdatatext valueattrs=graphdatatext;
blockplot x=x block=std /
display=(values label outline) valuehalign=center
labelattrs=graphdatatext valueattrs=graphdatatext;
blockplot x=x block=mean /
Generalized Macro for BOXPLOTPARM Data 1361

display=(values label outline) valuehalign=center

labelattrs=graphdatatext valueattrs=graphdatatext;
blockplot x=x block=median /
display=(values label outline) valuehalign=center
labelattrs=graphdatatext valueattrs=graphdatatext;
endInnerMargin;
boxplotparm y=value x=x stat=stat /
datalabel=datalabel spread=true;
endlayout;
endgraph;
end;
run;

proc sgrender data=boxdata2 template=boxplotparm2;

run;
1362 Appendix 5 • Generalized Macro for BOXPLOTPARM Data
1363

Appendix 6
Memory Management for ODS
Graphics

SAS Options Affecting Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363

Managing a Java Out of Memory Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363

SAS Options Affecting Memory

ODS Graphics uses Java technology to produce its graphs. Most of the time this fact is
transparent to you because the required Java Runtime Environment (JRE) and JAR files
are included with SAS software installation. Also, the Java environment is automatically
started and stopped for you. When Java is started, it allocates a fixed amount of memory.
The memory can grow up to the value set for the -Xmx suboption in the JREOPTIONS
option (discussed in a moment). This memory is independent of the memory limit that
SAS sets for the SAS session with its MEMSIZE= option.
Normally, the memory limit for Java is sufficient for most ODS Graphics applications.
However, some tasks are very memory intensive and might exhaust all available Java
memory, resulting in an OutOfMemoryError condition. You might encounter Java
memory limitations in the following cases:
• the product of the output size and the DPI setting results in very large output
• a classification panel has a very large number of classifier crossings
• a scatter plot matrix has a large number of variables
• creating 3-D plots and 2-D contours, which are memory intensive to generate
• a plot has a very large number of marker labels
• a plot uses many character variables or has a large number of GROUP values
• using the SG Editor to edit a graph with a large amount of data

Managing a Java Out of Memory Error

If you encounter a Java OutOfMemoryError, then you can try executing your program
again by restarting SAS and specifying a larger amount of memory for Java at SAS
invocation.
1364 Appendix 6 • Memory Management for ODS Graphics

To determine what the current Java memory settings are, you can submit a PROC
OPTIONS statement that shows the value of the JREOPTIONS option:

proc options option=jreoptions;

run;

After you submit this procedure code, a list of JREOPTIONS settings is written to the
SAS log. The JREOPTIONS option has many suboptions that configure the SAS Java
environment. Many of the suboptions are installation and host specific and should not be
modified, especially the ones that provide installed file locations. For managing memory,
look for the -Xmx and -Xms suboptions:

JREOPTIONS=(/* other Java suboptions */ -Xmx128m -Xms128m)

-Xms
Use this option to set the minimum Java memory (heap) size, in bytes. Set this value
to a multiple of 1024 greater than 1MB. Append the letter k or K to indicate
kilobytes, or m or M to indicate megabytes. The default is 2MB. Examples:

-Xms6291456
-Xms6144k
-Xms6m

-Xmx
Use this option to set the maximum size, in bytes, of the memory allocation pool. Set
this value to a multiple of 1024 greater than 2MB. Append the letter k or K to
indicate kilobytes, or m or M to indicate megabytes. The default is 64MB. Examples:

-Xmx83886080
-Xmx81920k
-Xmx80m

As a general rule, you should set the minimum heap size (-Xms) equal to the maximum
heap size (-Xmx) to minimize garbage collections.
Typically, SAS sets both -Xms and -Xmx to be about 1/4 of the total available memory
or a maximum of 128M. However, you can set a more aggressive maximum memory
(heap) size, but it should never be more than 1/2 of physical memory.
You should be aware of the maximum amount of physical memory your computer has
available. Let us assume that doubling the Java memory allocation is feasible. So when
you start SAS from a system prompt, you can add the following option:

-jreoptions (-Xmx256m -Xms256m)

Alternatively, you might need to specify the setting in quotation marks:

-jreoptions '(-Xmx256m -Xms256m)'

The exact syntax varies for specifying Java options, depending on your operating
system, and the amount of memory that you can allocate varies from system to system.
The set of JRE options must be enclosed in parentheses. If you specify multiple
JREOPTIONS system options, then SAS appends JRE options to JRE options that are
currently defined. Incorrect JRE options are ignored.
Managing a Java Out of Memory Error 1365

If you choose to create a custom configuration file, then you would simply replace the
existing -Xms and -Xmx suboption values in the JREOPTIONS=(all Java options)
portion of the configuration file.
For more information, see the SAS Companion for your operating system.
1366 Appendix 6 • Memory Management for ODS Graphics
1367

Recommended Reading

Here is the recommended reading list for this title:

• SAS Graph Template Language: User's Guide
• Getting Started with the Graph Template Language in SAS: Examples, Tips, and
Techniques for Creating Custom Graphs
• PROC TEMPLATE Made Easy: A Guide for SAS(R) Users
• Statistical Graphics in SAS: An Introduction to the Graph Template Language and
the Statistical Graphics Procedures
• Statistical Graphics Procedures by Example: Effective Graphs Using SAS

For a complete list of SAS publications, go to sas.com/store/books. If you have

questions about which titles you need, please contact a SAS Representative:
SAS Books
SAS Campus Drive
Cary, NC 27513-2414
Phone: 1-800-727-0025
Fax: 1-919-677-4444
Email: [email protected]
Web address: sas.com/store/books
1368 Recommended Reading
1369

Glossary

anti-aliasing
a rendering technique for improving the appearance of text and curved lines in a
graph by blurring the jagged edges normally present. The degree of improvement is
relative to the nature of the graphical content (for example, vertical and horizontal
lines do not benefit from anti-aliasing). Extra processing is required to perform anti-
aliasing.

attribute bundle
a common collection of visual properties associated with a graphical primitive such
as a line, marker, or text. For example, all lines have visual properties of pattern,
thickness, and color. All markers have visual properties of symbol, size, weight, and
color. Attribute bundles can be associated with style elements in order to indirectly
assign visual properties.

axis
a line that represents the midpoints (for a discrete axis) or the scale (for a continuous
or interval axis) for graphing variable or data values. An axis typically consists of an
axis line with tick marks, tick values (or midpoint values), and a label.

axis offset
the gaps that normally appear at the ends of an axis line. The gaps enable markers,
bars, and other graphic primitives that are drawn at extreme data values to be
rendered without clipping. An offset can also be used to add extra space between an
axis line and visual elements in the graph.

axis threshold
a numerical bias from 0 to 1 that determines whether an extra tick is added at either
end of a non-discrete, interval axis. If the minimum and maximum thresholds are set
to 0, then no ticks are added beyond the actual data range. If both minimum and
maximum thresholds are set to 1, then the data range is completely bounded by the
first and last ticks.

axis tick mark

a short line segment perpendicular to the axis line. A tick can cross the axis line, or
be drawn from the axis inside or outside the wall.

axis tick value

a formatted data value represented by a tick mark.
1370 Glossary

axis type
a keyword that denotes axis functionality. For example, the axis type of interval axes
can be LINEAR, TIME, or LOG. The axis type of a discrete axis is DISCRETE.

band plot
a plot that draws a horizontal band with two Y values for each X value, or that draws
a vertical band with two X values for each Y value. A band plot is typically used to
show confidence, error, prediction, or control limits. The points on the upper and
lower band boundaries can be joined to create two outlines, or the area between the
boundaries can be filled.

bin
one of multiple numeric intervals into which continuous numeric data can be
categorized.

binned data
data that has been summarized or transformed in some way to facilitate its rendering
by a parameterized plot. Continuous numeric data is typically binned by setting a bin
width (interval size) and then computing the number of bins, or by setting the
number of bins and computing the bin width.

block
See statement block.

block plot
a plot that displays one or more rectangles (blocks) along an axis, where each
rectangle identifies a block of consecutive observations having the same value for a
specified block variable.

category variable
a classification variable with a finite number of distinct (discrete) values. These
variables are typically used to split data into subsets. For example, in a bar chart,
each unique value is displayed as a bar on a DISCRETE axis. In another example,
the variable payment mode can have two values, prepaid and postpaid. Customers
can be classified based on this variable as prepaid customers and postpaid customers.

cell
See graph cell.

cell block
a block beginning with a CELL statement and ending with an ENDCELL statement
that defines the graphical content of a cell. The cell block is available only within a
LATTICE layout.

child block
a block that is contained within another block when two or more blocks are nested.
For example, a CELLHEADER block is always a child of a CELL block.

class variable
See classification variable.

classification level
for a single classification variable, each unique value is regarded as a classification
level. For two or more variables, a classification level is one of the unique
combinations (crossings) of the unique values of each variable. For example, if three
variables have four, two, and three distinct values, there are 24 classification levels.
Glossary 1371

classification panel
a multi-cell graph in which the cell data is driven by the values of one or more
classification variables. The number of the cells is determined by the unique values
of the classification variables. Each cell of the panel has the same types of plots.

classification variable (class variable)

a variable whose values are used to classify the observations in a data set into
different groups that are meaningful for analysis. A classification variable can have
either character or numeric values. Classification variables include group, subgroup,
category, and BY variables.

clip
to truncate a plot or graphical element (such as a line, marker, or band) when it
reaches a boundary such as a plot wall.

column axis
an external axis appearing above or below a column of cells and serving as a
common reference for the column of a multi-cell layout, such as a LATTICE,
DATAPANEL, or DATALATTICE layout.

column gutter
the space between columns of cells in a multi-cell layout.

column header
text that labels the column contents in a multi-cell layout. This text can be aligned
above or below the cells in a column. In a LATTICE layout, the column header is not
restricted to text (it can contain a plot or a legend, for example).

column major order

an order for populating cells of a layout or entries in a legend when the number of
rows is specified. By default, cells or entries are filled starting from the top left and
moving down. When the bottom row of the first column is filled, a new column
begins filling to the right of the previous column, and so on until all content items
have been placed in cells or entries. There might be empty cells or entries in the last
column.

column weight
in a LATTICE layout, the proportion of width allotted to a specific column of the
layout. The sum of all column weights is 1.

computed plot
a plot in which input data is internally summarized or otherwise transformed to
create new data that is actually rendered by the plot. Examples of computed plot
statements are BARCHART, BOXPLOT, HISTOGRAM, ELLIPSE, and
REGRESSIONPLOT.

conditional logic
syntax that enables one set of statements or an optional alternate set of statements to
execute at run time.

continuous legend
a legend that shows a mapping between a color ramp or color segments and
corresponding numeric values. Plots that support a COLORMODEL= option can use
this type of legend.
1372 Glossary

crossing
a combination of the unique values of one or more classification variables.

cube
in three-dimensional graphics, the outlines formed by the intersection of three pairs
of parallel planes; each pair is orthogonal to the primary X, Y, and Z axes. The
display of the cube is optional.

data object
a transient version of a SAS data set created by ODS. When an input SAS data set is
bound to a compiled graph template, an ODS data object is created, based on all the
columns requested in the template definition and any new columns that have been
directly or indirectly computed. A data object can persist when used with the ODS
OUTPUT statement.

data tip
data or other detailed information that is displayed when a user positions a mouse
pointer over an element in a graph. For example, a data tip typically displays the data
value that is represented by a bar, a plot point, or some other element.

define block
in the TEMPLATE procedure, a define block (beginning with a DEFINE statement
and ending with an END statement) creates various types of templates, including
STATGRAPH, STYLE, and TABLE.

dependent plot
a plot that cannot be rendered by itself. Dependent plots must be overlaid with a
stand-alone plot. Dependent plots do not provide data ranges to establish axes.
REFERENCELINE, DROPLINE, and LINEPARM statements produce dependent
plots.

design size
the intended size of a graph that is specified in the graph template definition. The
DESIGNHEIGHT and DESIGNWIDTH options of the BEGINGRAPH statement
set the intended height and width, which are used to determine the scale factors when
the graph is resized. The intended height and width are used unless overridden by the
ODS Graphics statement HEIGHT or WIDTH options when the template is
executed.

device-based graphic
a graph created with SAS/GRAPH software for which a user-specified or default
device (DEVICE= option) controls certain aspects of the graphical output.

discrete axis
an axis for categorical data values. The distance between ticks has no significance. A
bar chart always has a discrete axis.

discrete legend
a legend that provides values or descriptive information about graphical elements in
a grouped or overlaid plot.

dots per inch (DPI)

a measure of the graph resolution by its dot density.

DPI
See dots per inch.
Glossary 1373

drop line
a line drawn from a point in the plot area perpendicular to an axis.

dynamic variable
a variable defined in a template with the DYNAMIC statement that can be initialized
at template run time.

equated axes
in two-dimensional plots, axes that use the same drawing scale (ratio of display
distance to data interval) on both axes. For example, an interval of 2 on the X axis
maps to the same display distance as an interval of 2 on the Y axis. The aspect ratio
of the plot display equals the aspect ratio of the plot data. In other words, a 45-degree
slope in data will be represented by a 45-degree slope in the display. Equated axes
are always of TYPE=LINEAR. The number of intervals displayed on each axis does
not have to be the same.

external axis
an axis that is outside all cells of a layout. An external axis represents a common
scale for all plots in a row or column of a multi-cell layout.

fill
to apply a color within a bounded area. Many plots, such bar charts and band plots,
have bounded areas that can be filled or unfilled. When filled, a color is applied.
When unfilled, the areas are transparent.

fit policy
one of several algorithms for avoiding tick-value collision when space allotted to a
predefined area does not permit all the text to fit. For example, an axis might have a
THIN policy that eliminates the display of tick values for alternate ticks. A ROTATE
policy would turn the tick values at a 45-degree angle. A TRUNCATE policy would
truncate all long tick values to a fixed length and add an ellipsis (. . .) at the end to
imply truncation. A STAGGER policy would create two rows of tick values with
consecutive tick values alternating between rows. A compound policy such as
STAGGERROTATE could be used to automatically choose the best fit policy for the
situation.

footnote area
the region below the graph area where text produced by ENTRYFOOTNOTE
statements appears.

frequency variable
in an input data set, a non-negative and non-zero integer variable that represents the
frequency of occurrence of the current observation, essentially treating the data set as
if each observation appeared n times, where n is the value of the FREQ variable for
the observation.

fringe plot
a plot consisting of short, equal-length line segments drawn from and perpendicular
to an axis. Each observation of a numeric variable corresponds to the location for a
line segment.

function
See SAS function.
1374 Glossary

glyph
the most basic element (a grapheme or combination of graphemes) of a typeface or
font that carries meaning in the text of a writing system. For example, the Z character
can be represented by a number of different glyphs--boldface, italic, or in varying
font styles, all of which represent the the letter "Z."

graph cell (cell)

a distinct rectangular subregion of a graph that can contain plots, text, or legends.

graph panel
a graph with multiple cells.

graphics template
See ODS template.

grid
a uniform arrangement of the rows and columns of a multi-cell layout.

gridded data
input that contains at least three numeric variables. Two of the variables are treated
as X and Y variables and the third variable Z is treated as if it were a function of X
and Y. The X and Y variable values occur at uniformly spaced intervals (although the
size and number of intervals might be different for X and Y). All X,Y pairs are
unique, and Z values are interpolated so that every X,Y pair has a Z value. Raw data
that has at least three numeric variables can be converted to gridded data with the
G3GRID procedure (in SAS/GRAPH). The procedure offers both bivariate and
spline interpolation methods for computing Z values.

group index
a numeric variable with positive integer values that correspond to values of a group
variable. The index values are used to associate GraphData1 GraphDataN style
elements with group values.

group variable
a variable in the input data set that is used to categorize chart variable values into
groups. A group variable enables the data for each distinct group value to be
rendered in a visually different manner. For example, a grouped scatter plot displays
a distinct marker and color for each group value.

image format
a file format that displays a graphical representation. PNG, GIF, TIFF, and JPEG are
examples of image formats, each with different characteristics.

inset
a graphical element such as a legend, line of text, or a table of text that is embedded
inside of a graph's plot area.

interval axis
an axis where the distance between tick marks represents monotonically increasing
or decreasing numeric units of some scale (like a ruler). The standard interval axis is
called a LINEAR axis. Specialized interval axes include a TIME axis and a LOG
axis.

layout
a generic term for a rectangular container that lays out the positions and sizes of its
child components.
Glossary 1375

layout block
a block beginning with a LAYOUT statement and ending with an ENDLAYOUT
statement.

layout grid
a multi-cell layout arranged as a grid of cells in rows and columns.

layout row (row)

a set of layout cells that are side-by-side and share the same alignment.

layout type
a keyword indicating the functionality of the layout. For example OVERLAY,
LATTICE, and DATAPANEL are layout types.

legend entry
a combination of a graphical element such as a marker or line along with text
describing the value or use of the graphical element. A discrete legend can have
several legend entries.

legend title
text that explains how to interpret the legend.

line property
a value that defines the pattern, thickness, or color of a line. By default, the value for
a line property is derived from a style element in the current style.

linear axis
an interval axis with ticks placed on a linear scale.

loess plot
a curved line showing a loess fit for a set of points.

log axis
an axis displaying a logarithmic scale. A log axis is useful when data values span
orders of magnitude.

macro variable reference

a string that contains the name of a macro variable that is referenced in order to
substitute a value that is located or defined elsewhere.

marker
a symbol such as a diamond, a circle, or a triangle that is used to indicate the location
of, or annotate, a data point in a plot or graph.

marker property
a value that defines the symbol used as a marker, or its size, weight, or color. By
default, the value for a marker property is derived from a style element in the current
style.

multi-cell layout
a layout that supports a rectangular grid of cells, each of which can contain a
graphical element, such as a plot, a legend, a nested layout, and so on.

nested layout
a layout block that appears within the scope of another layout block.
1376 Glossary

ODS
See Output Delivery System.

ODS Graphics
an extension to ODS that is used to create analytical graphs using the Graph
Template Language.

ODS Graphics Editor

an interactive application that can be used to edit and annotate ODS Graphics output.

ODS style (style)

a combination of colors, fonts, lines, marker symbols, and so on that provide a
specific appearance for SAS output. A style is defined in ODS by a style template.

ODS template (graphics template)

a description of how output should appear when it is formatted. ODS templates are
stored as compiled entries in a template store, also known as an item store. Common
template types include STATGRAPH, STYLE, CROSSTABS, TAGSET, and
TABLE.

opaque
a property of a background. Opaque backgrounds are filled with a color. Non-opaque
backgrounds are transparent.

outlier
a data point that differs from the general trend of the data by more than is expected
by chance alone. An outlier might be an erroneous data point or one that is not from
the same sampling model as the rest of the data.

Output Delivery System (ODS)

a component of SAS software that can produce output in a variety of formats such as
markup languages (HTML, XML), PDF, listing, RTF, PostScript, and SAS data sets.

overlay
a plot that can be superimposed on another plot when specified within an overlay-
type layout. A common overlay combination is a fit line on a scatter plot.

overlay layout
a type of layout that supports the superimposition of graphical components, such as
plots, legends, and nested layouts.

parameterized plot
a non-computed plot that requires parameterized data. The Graph Template
Language offers several plots in both computed and parameterized versions, for
example, BARCHART and BARCHARTPARM. Some computed plots such as
REGRESSIONPLOT can be emulated with a SERIESPLOT if the input data
represented points on a fit line.

parent block
when two or more blocks are nested, any layout block that contains one or more
layout blocks is a parent of the contained blocks.

plot
a visual representation of data such as a scatter plot, needle plot, or contour plot.
Glossary 1377

plot area
the space, bounded by the axes, where a visual representation of data, such as a
scatter plot, a series line, or a histogram, is drawn.

plot type
a plot family such as bar chart (which would include horizontal, vertical, and
grouped bar charts), or a classification scheme for plots based on some useful
criteria, such as whether the plots are computed or parameterized.

primary axis
the X or Y axis contrasted to the X2 or Y2 secondary axis.

primary plot
the plot in an overlay that determines axis features, such as axis type and axis label.

prototype layout
an overlay plot composite that appears in each cell of a classification panel. Each
instance of the prototype represents a different subset (classification level) of the
data.

regression plot
a straight or curved line showing a linear or higher order regression fit for a set of
points.

required argument
a variable or constant that must be specified in order to evaluate an expression or
render a plot, legend, text, or a layout. For example, a scatter plot has two required
arguments: X=column and Y=column.

role
a description of the purpose that a variable serves in a plot. For example, a series plot
has predefined roles named for X , Y, GROUP, and CURVELABEL.

row
See layout row.

row axis
an external axis appearing on the left or right of a row of cells in a multi-cell layout.

row gutter
space between rows of cells of a multi-cell layout.

row header
typically, the text that identifies the row contents in a multi-cell layout. This text can
be aligned to the right or left of the cells in a row. The row header is not restricted to
text (it can contain a plot or a legend, for example).

row major order

an order for populating cells of a layout or entries of a legend when the number of
columns is specified. For example, in the default case: Start at the top left and fill
cells or entries left-to-right. When the right-most column is filled, begin a new row
below the previous row. Continue this until all content items have been placed in
cells or entries. There might be empty cells/entries in the last row.
1378 Glossary

row weight
in a LATTICE layout, the proportion of height allotted to a specific row of the
layout. The sum of all row weights is 1.

SAS function (function)

a type of SAS language element that is used to process one or more arguments and
then to return a result that can be used in either an assignment statement or an
expression.

secondary axis
an X2 or Y2 axis, as contrasted with the primary axes X or Y.

SGE file
a file created in the ODS Graphics environment that contains an editable graph. Such
files have a .sge file extension and can be edited only with the ODS Graphics Editor.
You can edit SGE files from the SAS Results window or by opening the SGE file
from within the ODS Graphics Editor.

sidebar
an area of certain multi-cell layouts external to the grid of cells where text or other
graphical elements can appear. The LATTICE, DATAPANEL, and DATALATTICE
layout support four sidebar areas (TOP, BOTTOM, LEFT, and RIGHT).

single-cell layout
a layout type that supports only one cell. The OVERLAY, OVERLAY3D, and
OVERLAYEQUATED layouts are examples of single-cell layouts.

sparse data
in classification panels with two or more classifiers, some crossings of the
classification values might not be present in the input data. Such input data is called
sparse data. By default, a DATAPANEL layout does not generate cells for sparse
data, but if requested, it can produce empty cells as place holders for the non-existent
crossings.

stand-alone plot
a plot that has its own data range and can therefore appear by itself in a layout.

statement block (block)

a group of statements that has both a logical beginning and ending statement. For
example, a LAYOUT statement along with its ENDLAYOUT statement and all
contained statements are a block. Some blocks can be nested within other blocks.

style
See ODS style.

style attribute
a visual property, such as color, font properties, and line characteristics, that is
defined in ODS with a reserved name and value. Style attributes are collectively
referenced by a style element within a style template.

style element
a named collection of style attributes that affects specific parts of ODS output. For
example, a style element might specify the color and font properties of title text or
other text in in a table or graph.
Glossary 1379

style reference
a part of the Graph Template Language syntax that indicates the current value of a
specific attribute of a specific style element. For example,
SIZE=GraphTitleText:FontSize means to assign to SIZE the value of the FontSize
attribute of the GraphTitleText style element from the current style.

surface plot
a three-dimensional graph that displays values of a vertical Z variable based on
gridded X and Y variables.

template compile time

the phase when the source program of a template definition is submitted. The syntax
of the definition is evaluated for correctness. If no errors are detected, the definition
is converted to a binary format and stored for later access.

template definition (template source)

the TEMPLATE procedure source program that creates a template. A template
definition can be generated from a compiled template. Also called the template
source.

template run time

the actions performed when a compiled template is bound to a data object and then
rendered to produce a graph. Run-time errors can occur that prevent a graph from
being produced.

template source
See template definition.

template store
an item store that contains definitions that were created by the TEMPLATE
procedure. Definitions that SAS provides are in the item store Sashelp.Tmplmst. You
can store definitions that you create in any template store to which you have Write
access.

template-based graphic
graphical output produced by a compiled ODS template of the type STATGRAPH.
That is, a graph that is produced within the ODS graphics environment rather than in
the traditional device-based environment.

text property
any of a common set of characteristics that can be specified for any text string: color,
family, size, weight, and style. By default, values for these properties are derived
from a style element in the current style.

threshold range
a lower and upper value that specifies the first and last tick mark values for a
continuous, linear axis.

time axis
an axis type that displays only SAS date, time, or datetime values. Axis tick value
increments can be specified as time or date intervals, such as MINUTE, HOUR,
DAY, WEEK, MONTH, QUARTER, or YEAR.

title area
the region above the graph area where text produced by ENTRYTITLE statements
appears.
1380 Glossary

transparency
the degree to which a graphic element (such as a marker or filled area) is opaque or
transparent. Transparency is indicated with a number from 0 (completely opaque) to
1 (completely transparent).

Unicode
a 16-bit encoding that is the industry standard for supporting the interchange,
processing, and display of characters and symbols from most of the world's writing
systems.

wall
the area bounded by orthogonal axis pairs. In two-dimensional graphs, there is one
wall bounded by the XY axes. In three-dimensional graphs, there are three walls,
bounded by the XY, YZ, and XZ axes. A wall has an optional outline and can be
opaque or transparent.

weight variable
a numeric variable that represents a weight (for example, costs) to be applied to
observations.
1381

Index

A STEPPLOT statement 778

ACROSS= option ARROWHEADS= option
AXISLEGEND statement 1091 VECTORPLOT statement 840
DISCRETELEGEND statement 1111 ARROWHEADSCALE= option
MERGEDLEGEND statement 1111 DRAWARROW statement 1215
ADDITIONALNAMES= option SERIESPLOT statement 744
MERGEDLEGEND statement 1112 STEPPLOT statement 778
ALIGN= ARROWHEADSHAPE= option
INNERMARGIN statement 166 DRAWARROW statement 1214
SIDEBAR statement, datalattice layout SERIESPLOT statement 745
65 STEPPLOT statement 778
SIDEBAR statement, datapanel layout VECTORPLOT statement 840
92 ASORT function 1325
SIDEBAR statement, lattice layout 134 ASPECTRATIO= option
ALPHA= option LAYOUT OVERLAY statement 137
ELLIPSE statement 424 LAYOUT PROTOTYPE statement 160
ALPHA= regression option attribute maps
LOESSPLOT statement 563 for discrete values 1287
PBSPLINEPLOT statement 610 for numeric ranges 1301, 1306
REGRESSIONPLOT statement 676 ATTRMAP= argument
ALTDISPLAY= option DISCRETEATTRVAR statement 1298
datalattice, datapanel axis options 1034 RANGEATTRVAR statement 1309
ALTDISPLAYSECONDARY= option ATTRVAR= argument
datalattice, datapanel axis options 1034 DISCRETEATTRVAR statement 1297
ALTFILLATTRS= option RANGEATTRVAR statement 1308
BLOCKPLOT statement 291 AUTOALIGN= option
anatomy of a graph 7 AXISLEGEND statement 1091
ANCHOR= option CONTINUOUSLEGEND statement
DRAWIMAGE statement 1221 1100
DRAWOVAL statement 1235 DISCRETELEGEND statement 1112
DRAWRECTANGLE statement 1243 ENTRY statement 1148
DRAWTEXT statement 1251 LAYOUT GRIDDED statement 103
ANNOTATE statement LAYOUT LATTICE statement 113
TEMPLATE procedure 1267 MERGEDLEGEND statement 1112
area fill AUTOITEMSIZE= option
remapping in grouped data 183 DISCRETELEGEND statement 1113
arrow MERGEDLEGEND statement 1113
drawing an arrow 1212, 1219 axis features
ARROWDIRECTION= option axis types 883
VECTORPLOT statement 840 COLUMN2AXES block 879
ARROWHEADDIRECTION= option COLUMN2DATARANGE= 877
DRAWARROW statement 1214 COLUMNAXES block 879
ARROWHEADPOSITION= option COLUMNDATARANGE= 877
SERIESPLOT statement 744 controlling 883
1382 Index

data mapping 876 CONTINUOUSLEGEND statement

displayed axis features 875 1100
in data lattice layouts 882 DISCRETELEGEND statement 1113
in data panel layouts 882 ENTRY statement 1149
in gridded layouts 876, 883 ENTRYFOOTNOTE statement 1156
in lattice-type layouts 877, 882 ENTRYTITLE statement 1163
in layouts 875, 876 LAYOUT DATALATTICE statement
in overlay-type layouts 881 48
OFFSETMAX= 886 LAYOUT DATAPANEL statement 74
OFFSETMIN= 886 LAYOUT GRIDDED statement 104
overview 10, 875 LAYOUT LATTICE statement 114
ROW2AXES block 879 LAYOUT OVERLAY statement 137
ROW2DATARANGE= 877 LAYOUT OVERLAY3D statement 154
ROWAXES block 879 LAYOUT OVERLAYEQUATED
ROWDATARANGE= 877 statement 145
setting axis types 884 LAYOUT REGION statement 162
THRESHOLDMAX= 885 MERGEDLEGEND statement 1113
THRESHOLDMIN= 885 BACKLIGHT= option
VIEWMAX= 884 POLYGONPLOT statement 631
VIEWMIN= 884 TEXTPLOT statement 814
when plots share data 880 BANDPLOT statement
axis options TEMPLATE procedure 204
LAYOUT DATALATTICE statement BARCHART statement
1032 TEMPLATE procedure 218
LAYOUT DATAPANEL statement BARCHARTPARM statement
1032 TEMPLATE procedure 250
LAYOUT LATTICE statement 963 BARLABEL= option
LAYOUT OVERLAY statement 889 BARCHART statement 221
LAYOUT OVERLAY3D statement 945 WATERFALLCHART statement 857
LAYOUT OVERLAYEQUATED BARLABELATTRS= option
statement 1013 BARCHART statement 221
axis statements in LATTICE layout WATERFALLCHART statement 857
COLUMN2HEADERS statement 133 BARLABELFITPOLICY= option
COLUMNHEADERS statement 133 BARCHART statement 221
ROW2HEADERS statement 133 WATERFALLCHART statement 857
ROWHEADERS statement 133 BARLABELFORMAT= option
axis, defined 7 BARCHART statement 222
AXISBREAKSYMBOL= option WATERFALLCHART statement 858
BEGINGRAPH statement 23 BARWIDTH= option
AXISBREAKTYPE= option BARCHART statement 222
BEGINGRAPH statement 24 BARCHARTPARM statement 253
AXISLEGEND statement HIGHLOWPLOT statement 474
TEMPLATE procedure 1089 WATERFALLCHART statement 858
AXISLINEEXTENT= option BASE=
BEGINGRAPH statement 25 LOGOPTS= option, datalattice or
AXISTABLE statement datapanel axis 1066
TEMPLATE procedure 190 LOGOPTS= option, lattice axis 994
LOGOPTS= option, overlay axis 926
BASELINEATTRS= option
B BARCHART statement 223
BACKGROUNDCOLOR= argument BARCHARTPARM statement 253
INNERMARGIN statemen 167 NEEDLEPLOT statement 587
BACKGROUNDCOLOR= option WATERFALLCHART statement 858
AXISLEGEND statement 1091 BASELINEINTERCEPT= option
BEGINGRAPH statement 26 BARCHART statement 223
BARCHARTPARM statement 254
Index 1383

LINECHART statement 524 CONTINUOUSLEGEND statement

NEEDLEPLOT statement 587 1101
WATERFALLCHART statement 859 DISCRETELEGEND statement 1113
BEGINGRAPH statement 4 DRAWIMAGE statement 1221
DESIGNHEIGHT= option 706 DRAWTEXT statement 1251
TEMPLATE procedure 21 ENTRY statement 1149
BEGINPOLYGON statement ENTRYFOOTNOTE statement 1156
TEMPLATE procedure 1199 ENTRYTITLE statement 1164
BEGINPOLYLINE statement LAYOUT DATALATTICE statement
TEMPLATE procedure 1206 49
BIHISTOGRAM3DPARM statement LAYOUT DATAPANEL statement 74
TEMPLATE procedure 284 LAYOUT GLOBALLEGEND
BINAXIS= option statement 97
BIHISTOGRAM3DPARM statement LAYOUT GRIDDED statement 104
285 LAYOUT LATTICE statement 114
HISTOGRAM statement 495 LAYOUT OVERLAY statement 138
HISTOGRAMPARM statement 508 LAYOUT OVERLAY3D statement 154
BINSTART= option LAYOUT OVERLAYEQUATED
HISTOGRAM statement 495 statement 146
BINWIDTH= option LAYOUT REGION statement 163
HISTOGRAM statement 496 MERGEDLEGEND statement 1113
BLOCK= argument BOUNDARY= option
BLOCKPLOT statement 291 HISTOGRAM statement 496
BLOCKINDEX= option BOXPLOT statement
BLOCKPLOT statement 291 TEMPLATE procedure 305
BLOCKPLOT statement BOXPLOTPARM statement
TEMPLATE procedure 289 TEMPLATE procedure 336
border, specifying on a graph 16 BOXWIDTH= option
BORDER= option BOXPLOT statement 308
AXISLEGEND statement 1091 BOXPLOTPARM statement 341
BEGINGRAPH statement 26 BREAK= option
CONTINUOUSLEGEND statement LINECHART statement 524
1101 SERIESPLOT statement 745
DISCRETELEGEND statement 1113 STEPPLOT statement 779
DRAWIMAGE statement 1221 BUBBLEPLOT statement
DRAWTEXT statement 1251 TEMPLATE procedure 367
ENTRY statement 1149 BUBBLERADIUSMAX= option
ENTRYFOOTNOTE statement 1156 BUBBLEPLOT statement 370
ENTRYTITLE statement 1164 BUBBLERADIUSMIN= option
LAYOUT DATALATTICE statement BUBBLEPLOT statement 370
48
LAYOUT DATAPANEL statement 74
LAYOUT GLOBALLEGEND C
statement 97 CAPSHAPE= option
LAYOUT GRIDDED statement 104 BOXPLOT statement 308
LAYOUT LATTICE statement 114 BOXPLOTPARM statement 341
LAYOUT OVERLAY statement 138 CATEGORY= argument
LAYOUT OVERLAY3D statement 154 BARCHART statement 221
LAYOUT OVERLAYEQUATED BARCHARTPARM statement 253
statement 146 LINECHART statement 523
LAYOUT REGION statement 162 PIECHART statement 615
MERGEDLEGEND statement 1113 WATERFALLCHART statement 857
ODS GRAPHICS statement 16 CATEGORY= option
BORDERATTRS= option MOSAICPLOTPARM statement 574
AXISLEGEND statement 1092 CATEGORYDIRECTION= option
BEGINGRAPH statement 26 PIECHART statement 615
1384 Index

CELL block DENDROGRAM statement 396

in LAYOUT LATTICE 127 CLUSTERWIDTH= option
cell, defined 7 AXISTABLE statement 193
CELLHEADER statement BARCHARTPARM statement 254
in LAYOUT LATTICE 128 BOXPLOT statement 309
CELLHEIGHTMIN= option BOXPLOTPARM statement 342
LAYOUT DATALATTICE statement HIGHLOWPLOT statement 476
49 NEEDLEPLOT statement 587
LAYOUT DATAPANEL statement 74 SCATTERPLOT statement 685
CELLWIDTHMIN= option SERIESPLOT statement 746
LAYOUT DATALATTICE statement STEPPLOT statement 779
49 TEXTPLOT statement 816
LAYOUT DATAPANEL statement 75 COLC function 1325
CENTERFIRSTSLICE= option COLLABEL function 1325
PIECHART statement 615 COLN function 1325
CHAR= argument COLNAME function 1325
SYMBOLCHAR statement 1173 color ramp
CLASS= argument attribute map for 1301
AXISTABLE statement 192 color value type 1340
CLASS= option COLOR= attribute
AXISTABLE statement 192 for fills 1348
BLOCKPLOT statement 292 for lines 1349
CLASSDISPLAY= option for markers 1350
AXISTABLE statement 192 for text 1351
CLASSORDER= option COLORBANDS=
AXISTABLE statement 193 DISCRETEOPTS= option, overlay axis
CLASSVARS= argument 904
LAYOUT DATAPANEL statement 73 COLORBANDSATTRS=
CLI= regression option DISCRETEOPTS= option, overlay axis
PBSPLINEPLOT statement 610 905
REGRESSIONPLOT statement 676 COLORBYFREQ= option
CLIP= option BARCHART statement 224
DROPLINE statement 417 COLORGROUP= argument
ELLIPSE statement 424 HEATMAPPARM statement 461
ELLIPSEPARM statement 433 COLORGROUP= option
LINEPARM statement 544 AXISTABLE statement 193
REFERENCELINE statement 656 MOSAICPLOTPARM statement 575
VECTORPLOT statement 841 WATERFALLCHART statement 860
CLIPCAP= option COLORMODEL= option
HIGHLOWPLOT statement 474 BARCHART statement 224
CLIPCAPSHAPE= option BARCHARTPARM statement 255
HIGHLOWPLOT statement 476 BUBBLEPLOT statement 370
CLM= regression option CONTOURPLOTPARM statement 388
LOESSPLOT statement 563 HEATMAP statement 448
PBSPLINEPLOT statement 610 HEATMAPPARM statement 462
REGRESSIONPLOT statement 676 HIGHLOWPLOT statement 477
CLOSE= option LINECHART statement 525
HIGHLOWPLOT statement 476 MOSAICPLOTPARM statement 575
CLUSTERAXIS= option POLYGONPLOT statement 632
SCATTERPLOT statement 684 SCATTERPLOT statement 682
SERIESPLOT statement 745 SCATTERPLOTMATRIX statement
STEPPLOT statement 779 718
TEXTPLOT statement 816 SERIESPLOT statement 746
CLUSTERHEIGHT= argument SURFACEPLOTPARM statement 805
DENDROGRAM statement 396 TEXTPLOT statement 815
CLUSTERS= option VECTORPLOT statement 841
Index 1385

WATERFALLCHART statement 860 LAYOUT DATAPANEL statement 75

COLORRESPONSE= argument LAYOUT LATTICE statement 114,
HEATMAPPARM statement 462 130
COLORRESPONSE= option COLUMNGUTTER= option
BARCHART statement 225 LAYOUT DATALATTICE statement
BARCHARTPARM statement 256 51
BUBBLEPLOT statement 371 LAYOUT DATAPANEL statement 77
HEATMAP statement 449 LAYOUT GRIDDED statement 104
HIGHLOWPLOT statement 477 LAYOUT LATTICE statement 116
LINECHART statement 525 COLUMNHEADERS statement
MOSAICPLOTPARM statement 576 LAYOUT LATTICE statement 133
POLYGONPLOT statement 633 COLUMNHEADERS= option
SCATTERPLOT statement 683 LAYOUT DATALATTICE statement
SCATTERPLOTMATRIX statement 51
719 COLUMNS= option
SERIESPLOT statement 747 LAYOUT DATALATTICE statement
SURFACEPLOTPARM statement 805 51
TEXTPLOT statement 815 LAYOUT DATAPANEL statement 76
VECTORPLOT statement 842 LAYOUT GRIDDED statement 104
WATERFALLCHART statement 861 LAYOUT LATTICE statement 116
colors COLUMNVAR= argument
remapping in grouped data 183 LAYOUT DATALATTICE statement
with grouped data 179 48
with non-grouped data 177 COLUMNWEIGHT= option
COLORSTAT= option LAYOUT DATALATTICE statement
BARCHART statement 226 52
HEATMAP statement 449 COLUMNWEIGHT== option
WATERFALLCHART statement 862 LAYOUT DATAPANEL statement 77
column value type 1340 COLUMNWEIGHTS= option
COLUMN2AXES block LAYOUT LATTICE statement 116
in LAYOUT LATTICE 132 COMMONAXISOPTS= option
COLUMN2AXISOPTS= option LAYOUT OVERLAYEQUATED
LAYOUT DATALATTICE statement statement 146, 1013
49 computed plots 10
LAYOUT DATAPANEL statement 75 conditional logic
COLUMN2DATARANGE= option example uses 1333
LAYOUT DATALATTICE statement IF statement 14, 1333
50 overview 14
LAYOUT DATAPANEL statement 76 why used 1334
LAYOUT LATTICE statement 115, CONNECT= option
130 BOXPLOT statement 309
COLUMN2HEADERS statement WATERFALLCHART statement 862
LAYOUT LATTICE statement 133 CONNECTATTRS= option
COLUMNAXES block BARCHART statement 227
in LAYOUT LATTICE 132 BARCHARTPARM statement 257
COLUMNAXIS statement BOXPLOTPARM statement 343
in LAYOUT LATTICE 132 WATERFALLCHART statement 862
LAYOUT LATTICE statement 963 CONNECTDECREASINGATTRS=
COLUMNAXISOPTS= option option
LAYOUT DATALATTICE statement WATERFALLCHART statement 863
49, 1032 CONNECTINCREASINGATTRS=
LAYOUT DATAPANEL statement 75, option
1032 WATERFALLCHART statement 863
COLUMNDATARANGE= option CONNECTORDER= option
LAYOUT DATALATTICE statement BANDPLOT statement 207
50 SERIESPLOT statement 747
1386 Index

STEPPLOT statement 780 BANDPLOT statement 208

CONTINUOUSLEGEND statement DENSITYPLOT statement 404
LEGEND statement 1098 LINEPARM statement 545
CONTOURPLOTPARM statement LOESSPLOT statement 556
TEMPLATE procedure 386 MODELBAND statement 568
CONTOURTYPE= option PBSPLINEPLOT statement 603
CONTOURPLOTPARM statement 389 REFERENCELINE statement 658
CONTRIBUTEOFFSETS= option REGRESSIONPLOT statement 669
SCATTERPLOT statement 685, 817 SERIESPLOT statement 749
CORROPTS= option STEPPLOT statement 781
SCATTERPLOTMATRIX statement CURVELABELUPPER= option
720 BANDPLOT statement 208
COUNT= option MODELBAND statement 567
MOSAICPLOTPARM statement 575 CUT= option
CSS function 1328 DENDROGRAM statement 396
CUBE= option CUTOPTS= option
LAYOUT OVERLAY3D statement 154 DENDROGRAM statement 397
curve labels CV function 1328
location and position in templates 185 CYCLEATTRS= option
location relative to plot area 185 LAYOUT OVERLAY statement 138
position relative to curve line 186 LAYOUT OVERLAY3D statement 154
CURVELABEL= option LAYOUT OVERLAYEQUATED
DENSITYPLOT statement 404 statement 146
LINEPARM statement 544 LAYOUT PROTOTYPE statement 160
LOESSPLOT statement 555
PBSPLINEPLOT statement 602
REFERENCELINE statement 657 D
REGRESSIONPLOT statement 668 DATALABEL= option
SERIESPLOT statement 748 BARCHARTPARM statement 257
STEPPLOT statement 780 BOXPLOT statement 310
CURVELABELATTRS= option BOXPLOTPARM statement 343
BANDPLOT statement 207 BUBBLEPLOT statement 372
DENSITYPLOT statement 404 HISTOGRAMPARM statement 509
LINEPARM statement 545 NEEDLEPLOT statement 588
LOESSPLOT statement 555 SCATTERPLOT statement 686
MODELBAND statement 567 SCATTERPLOTMATRIX statement
PBSPLINEPLOT statement 602 721
REFERENCELINE statement 657 SERIESPLOT statement 752
REGRESSIONPLOT statement 668 STEPPLOT statement 785
SERIESPLOT statement 748 VECTORPLOT statement 842
STEPPLOT statement 780 DATALABELATTRS= option
CURVELABELLOCATION= option BARCHARTPARM statement 257
BANDPLOT statement 208 BOXPLOT statement 310
DENSITYPLOT statement 404 BOXPLOTPARM statement 343
LINEPARM statement 545 BUBBLEPLOT statement 372
LOESSPLOT statement 555 HISTOGRAMPARM statement 509
MODELBAND statement 568 NEEDLEPLOT statement 588
PBSPLINEPLOT statement 602 PIECHART statement 616
REFERENCELINE statement 657 SCATTERPLOT statement 687
REGRESSIONPLOT statement 669 SCATTERPLOTMATRIX statement
SERIESPLOT statement 748 721
STEPPLOT statement 781 SERIESPLOT statement 752
CURVELABELLOWER= option STEPPLOT statement 785
BANDPLOT statement 208 VECTORPLOT statement 842
MODELBAND statement 567 DATALABELCONTENT= option
CURVELABELPOSITION= option PIECHART statement 616
Index 1387

DATALABELFITPOLICY= option BIHISTOGRAM3DPARM statement

BARCHARTPARM statement 257 285
HISTOGRAMPARM statement 509 BLOCKPLOT statement 292
DATALABELLOCATION= option BOXPLOT statement 313
PIECHART statement 616 BOXPLOTPARM statement 346
DATALABELPOSITION= option BUBBLEPLOT statement 375
BUBBLEPLOT statement 372 DENDROGRAM statement 397
NEEDLEPLOT statement 588 DENSITYPLOT statement 408
SCATTERPLOT statement 687 DROPLINE statement 418
SCATTERPLOTMATRIX statement ELLIPSE statement 424
722 ELLIPSEPARM statement 433
SERIESPLOT statement 753 FRINGEPLOT statement 441
STEPPLOT statement 785 HEATMAP statement 450
VECTORPLOT statement 843 HEATMAPPARM statement 463
DATALABELSPLITCHAR= option HIGHLOWPLOT statement 479
HISTOGRAMPARM statement 510 HISTOGRAM statement 497
DATALABELSPLITCHARDROP= HISTOGRAMPARM statement 513
option LINECHART statement 527
HISTOGRAMPARM statement 511 LINEPARM statement 549
DATALABELSPLITJUSTIFY= option LOESSPLOT statement 559
BOXPLOT statement 312 MODELBAND statement 569
BOXPLOTPARM statement 345 MOSAICPLOTPARM statement 576
DATALABELTYPE= option NEEDLEPLOT statement 591
BARCHARTPARM statement 261 PBSPLINEPLOT statement 606
DATALATTICE layout 43, 45 PIECHART statement 618
DATAPANEL layout 43, 70 POLYGONPLOT statement 634
DATASKIN= option REFERENCELINE statement 661
BARCHART statement 227 REGRESSIONPLOT statement 673
BARCHARTPARM statement 262 SCATTERPLOT statement 690
BEGINGRAPH statement 27 SCATTERPLOTMATRIX statement
BOXPLOT statement 312 725
BOXPLOTPARM statement 345 SERIESPLOT statement 756
BUBBLEPLOT statement 375, 634 STEPPLOT statement 788
DROPLINE statement 418 SURFACEPLOTPARM statement 806
HIGHLOWPLOT statement 478 TEXTPLOT statement 817
HISTOGRAM statement 497 VECTORPLOT statement 846
HISTOGRAMPARM statement 512 WATERFALLCHART statement 864
LINECHART statement 526 DEFINE STATGRAPH statement 4
NEEDLEPLOT statement 591 DEGREE= regression option
PIECHART statement 617 LOESSPLOT statement 563
REFERENCELINE statement 661 PBSPLINEPLOT statement 610
SCATTERPLOT statement 690 REGRESSIONPLOT statement 677
SCATTERPLOTMATRIX statement DENDROGRAM statement
724 TEMPLATE procedure 395
SERIESPLOT statement 756 DENSITYPLOT statement
STEPPLOT statement 788 TEMPLATE procedure 402
VECTORPLOT statement 845 DESIGNHEIGHT= option
WATERFALLCHART statement 863 BEGINGRAPH statement 28
DATASYMBOLS= option BEGINGRAPH statement with
BEGINGRAPH statement 28 datalattice layout 67
DATATRANSPARENCY= option BEGINGRAPH statement with
and draw statements 1196 datapanel layout 73
AXISTABLE statement 194 DESIGNWIDTH= option
BANDPLOT statement 209 BEGINGRAPH statement 29
BARCHART statement 228 BEGINGRAPH statement with
BARCHARTPARM statement 262 datalattice layout 67
1388 Index

BEGINGRAPH statement with specifying fill color 1348

datapanel layout 73 specifying line attributes 177
DIAGONAL= option specifying line color 1349
SCATTERPLOTMATRIX statement specifying line pattern 1349
725 specifying line thickness 1349
dimension value type 1340 specifying marker attributes 178
DISCRETEATTRMAP statement specifying marker color 1350
TEMPLATE procedure 1287 specifying marker size 1350
DISCRETEATTRVAR statement specifying marker symbol 1350
TEMPLATE procedure 1297 specifying marker weight 1351
DISCRETEAXISOFFSETPAD= option specifying text color 1351
BEGINGRAPH statement 29 specifying text font family 1351
DISCRETELEGEND statement specifying text size 1351
TEMPLATE procedure 1109 specifying text style 1351
DISCRETELEGENDENTRYPOLICY= specifying text weight 1352
option specifying transparency 1348
DISCRETEATTRMAP statement 1288 DISPLAY=
DISCRETEMARKERSIZE= option XAXISOPTS=, YAXISOPTS= options,
SCATTERPLOT statement 690 overlayequated axis 1020
DISCRETEMAX= option DISPLAY= option
ODS GRAPHICS statement 396 and attribute options 184
DISCRETEOFFSET= option AXISTABLE statement 194
BARCHART statement 228 BANDPLOT statement 210
BARCHARTPARM statement 263 BARCHART statement 228
BEGINPOLYGON statement 1201 BARCHARTPARM statement 263
BEGINPOLYLINE statement 1208 BEGINPOLYGON statement 1201
BOXPLOT statement 313 BIHISTOGRAM3DPARM statement
BOXPLOTPARM statement 347 286
DRAWARROW statement 1215 BLOCKPLOT statement 292
DRAWIMAGE statement 1222 BOXPLOT statement 314
DRAWLINE statement 1229 BOXPLOTPARM statement 347
DRAWOVAL statement 1236 BUBBLEPLOT statement 376
DRAWRECTANGLE statement 1244 datalattice, datapanel axis options 1036
DRAWTEXT statement 1252 DRAWOVAL statement 1236
DROPLINE statement 419 DRAWRECTANGLE statement 1244
HIGHLOWPLOT statement 479 ELLIPSE statement 425
NEEDLEPLOT statement 592 ELLIPSEPARM statement 433
POLYGONPLOT statement 635 HEATMAP statement 450
REFERENCELINE statement 662 HEATMAPPARM statement 463
SCATTERPLOT statement 691 HIGHLOWPLOT statement 479
SERIESPLOT statement 756 HISTOGRAM statement 498
STEPPLOT statement 788 HISTOGRAMPARM statement 513
TEXTPLOT statement 818 lattice axis options 965
DISCRETEOPTS= option LINECHART statement 527
datalattice, datapanel axis options 1035 MODELBAND statement 569
lattice axis options 965 MOSAICPLOTPARM statement 576
overlay axis options 891 NEEDLEPLOT statement 592
DISCRETEX= option overlay axis options 891
HEATMAP statement 450, 463 overlay3d axis options 946
DISCRETEY= option PIECHART statement 619
HEATMAP statement 450, 463 POLYGONPLOT statement 635
display attributes SERIESPLOT statement 757
available line patterns 1352 STEPPLOT statement 789
cycling through group attributes 183 TEXTPLOT statement 818
grouped data 179 WATERFALLCHART statement 864
overview 177 DISPLAYCLIPPED= option
Index 1389

AXISLEGEND statement 1092 TEMPLATE procedure 416

DISCRETELEGEND statement 1113 DROPONMISSING= option
LAYOUT GLOBALLEGEND AXISTABLE statement 194
statement 97 DROPTO= option
MERGEDLEGEND statement 1113 DROPLINE statement 419
DISPLAYSECONDARY= DSORT function 1325
lattice axis options 131 DYNAMIC statement
XAXISOPTS=, YAXISOPTS= options, TEMPLATE procedure 13, 1313
overlayequated axis 1021 dynamic variables
DISPLAYSECONDARY= option See also dynamics
datalattice, datapanel axis options 1036 compared to macro variables 1315
lattice axis options 966 dynamics
overlay axis options 892 overview 13
DISPLAYSTATS= option using on templates 1313
BOXPLOT statement 315
BOXPLOTPARM statement 348
DISPLAYZEROLENGTHBAR= option E
BARCHART statement 229 ELLIPSE = option
BARCHARTPARM statement 264 SCATTERPLOTMATRIX statement
DOWN= option 726
AXISLEGEND statement 1092 ELLIPSE statement
DISCRETELEGEND statement 1114 TEMPLATE procedure 422
MERGEDLEGEND statement 1114 ELLIPSEPARM statement
draw statements TEMPLATE procedure 431
anchor position 1195 ENDLABELS= option
and the graph wall 1196 BIHISTOGRAM3DPARM statement
and the layout background 1196 286
drawing layers 1195 HISTOGRAM statement 498
drawing space 1192 HISTOGRAMPARM statement 514
drawing units 1192 ENTRY statement
example 1189 greek letters 1143, 1343
DRAWARROW statement prefix options 1138
TEMPLATE procedure 1212 reserved keywords 1142, 1343
DRAWIMAGE statement TEMPLATE procedure 1147
TEMPLATE procedure 1219 text commands 1140
DRAWLINE statement Unicode values 1142, 1343
TEMPLATE procedure 1227 ENTRYFOOTNOTE statement
DRAWORDER= option greek letters 1143, 1343
BUBBLEPLOT statement 376 prefix options 1138
DRAWOVAL statement reserved keywords 1142, 1343
TEMPLATE procedure 1233 TEMPLATE procedure 1154
DRAWRECTANGLE statement text commands 1140
TEMPLATE procedure 1241 Unicode values 1142, 1343
DRAWSPACE= option ENTRYTITLE statement
BEGINGRAPH statement 30 greek letters 1143, 1343
BEGINPOLYGON statement 1202 prefix options 1138
BEGINPOLYLINE statement 1208 reserved keywords 1142, 1343
DRAWARROW statement 1215 TEMPLATE procedure 1162
DRAWIMAGE statement 1222 text commands 1140
DRAWLINE statement 1229 Unicode values 1142, 1343
DRAWOVAL statement 1236 EQUATETYPE= option
DRAWRECTANGLE statement 1244 LAYOUT OVERLAYEQUATED
DRAWTEXT statement 1252 statement 146
DRAWTEXT statement ERRORBARATTRS= option
TEMPLATE procedure 1249 BARCHARTPARM statement 264
DROPLINE statement SCATTERPLOT statement 692
1390 Index

STEPPLOT statement 789 FILLATTRS= option

ERRORBARCAPSHAPE= option BANDPLOT statement 210
BARCHARTPARM statement 265 BARCHART statement 230
SCATTERPLOT statement 692 BARCHARTPARM statement 265
STEPPLOT statement 789 BEGINPOLYGON statement 1202
ERRORLOWER= option BIHISTOGRAM3DPARM statement
BARCHARTPARM statement 265 286
STEPPLOT statement 790 BLOCKPLOT statement 293
ERRORUPPER= option BOXPLOT statement 316
BARCHARTPARM statement 265 BOXPLOTPARM statement 350
STEPPLOT statement 790 BUBBLEPLOT statement 377
ESC keyword DRAWOVAL statement 1237
specifications inside quotation marks DRAWRECTANGLE statement 1245
1142 ELLIPSE statement 425
EVAL construct 12 ELLIPSEPARM statement 434
with expressions 1317 HEATMAP statement 451
with functions 1321 HEATMAPPARM statement 464
examples on the web 18 HIGHLOWPLOT statement 480
EXCLUDE= option HISTOGRAM statement 498
DISCRETELEGEND statement 1114 HISTOGRAMPARM statement 514
MERGEDLEGEND statement 1114 LEGENDITEM statement 1127
EXISTS function 1325 LINECHART statement 527
EXPAND function 1325 MODELBAND statement 570
expression value type 1341 MOSAICPLOTPARM statement 577
expressions PIECHART statement 619
and data filtering 1318 POLYGONPLOT statement 635
and data subsetting 1318 SURFACEPLOTPARM statement 806
and type conversion 1319 TEXTPLOT statement 819
difference from WHERE expressions VALUE statement 1290
1318 WATERFALLCHART statement 865
example uses 1317 FILLDISPLAY= option
on plot statement arguments 1318 LEGENDITEM statement 1127
overview 12 FILLEDOUTLINEDMARKERS= option
similarity to WHERE expressions 1318 LEGENDITEM statement 1128
why used 1317 LINECHART statement 528
EXTEND= option SCATTERPLOT statement 693
BANDPLOT statement 210 SERIESPLOT statement 757
LINEPARM statement 549 STEPPLOT statement 790
EXTRACTSCALE= option FILLITEMOUTLINE= option
CONTINUOUSLEGEND statement DISCRETELEGEND statement 1114
1101 FILLPATTERNATTRS= option
EXTRACTSCALETYPE= option BARCHART statement 230
CONTINUOUSLEGEND statement BARCHARTPARM statement 266
1101 HISTOGRAM statement 499
EXTREME= option HISTOGRAMPARM statement 515
BOXPLOT statement 316 FILLTYPE= option
BOXPLOTPARM statement 349 BARCHART statement 231, 267, 500,
515
BLOCKPLOT statement 294
F WATERFALLCHART statement 865
FAMILY= attribute FINALBARATTRS= option
for text 1351 WATERFALLCHART statement 865
fill attributes FINALBARTICKVALUE= option
remapping in grouped data 183 WATERFALLCHART statement 865
specifying fill color 1348 flexible templates, overview 12
specifying transparency 1348 fonts
Index 1391

See text attributes P99 1328

footnote area, defined 7 PROBT 1329
format value type 1341 Q1 1329
FORMAT= option Q3 1329
TEXTPLOT statement 819 QRANGE 1329
formats RANGE 1329
See SAS formats SKEWNESS 1329
Formats STDDEV 1329
user-defined, with Unicode values 1355 STDERR 1329
FREQ= option SUM 1329
BOXPLOT statement 317 summary statistics functions 1328
DENSITYPLOT statement 408 SUMWGT 1329
ELLIPSE statement 425 T 1329
HEATMAP statement 451 TYPEOF 1326
HISTOGRAM statement 500 types available 1321
SCATTERPLOT statement 694 UCLM 1329
SCATTERPLOTMATRIX statement USS 1329
727 VAR 1329
FREQ= regression option
PBSPLINEPLOT statement 611
REGRESSIONPLOT statement 677 G
FRINGEHEIGHT= option glyphs
FRINGEPLOT statement 441 lowercase Greek letters 1143, 1343
FRINGEPLOT statement special characters 1145, 1345
TEMPLATE procedure 439 uppercase Greek letters 1144, 1344
functions graph template language
ASORT 1325 See templates
COLC 1325 graph, defined 7
COLLABEL 1325 GraphBox style element 330, 331
COLN 1325 Displayopts attribute 331
COLNAME 1325 GraphBoxMean style element 330
CSS 1328 GraphBoxMedian style element 330
CV 1328 GraphBoxOutlier style element 330
DSORT 1325 GraphBoxWhisker style element 331
EXISTS 1325 graphics environment
EXPAND 1325 ODS GRAPHICS statement 176
GTL only functions 1324 GRIDATTRS=
IFC 1322 XAXISOPTS=, YAXISOPTS= options,
IFN 1322 overlayequated axis 1021
KURTOSIS 1328 GRIDATTRS= option
LCLM 1328 datalattice, datapanel axis options 1037
MAX 1328 lattice axis options 967
MEAN 1328 overlay axis options 893
MEDIAN 1328 overlay3d axis options 947
MIN 1328 GRIDDED layout 43, 102
N 1328 GRIDDED= option
NMISS 1328 CONTOURPLOTPARM statement 390
NUMERATE 1325 GRIDDISPLAY=
overview 12 XAXISOPTS=, YAXISOPTS= options,
P1 1328 overlayequated axis 1021
P25 1328 GRIDDISPLAY= option
P5 1328 datalattice, datapanel axis options 1038
P50 1328 lattice axis options 967
P75 1328 overlay axis options 893
P90 1328 overlay3d axis options 947
P95 1328 GROUP= option
1392 Index

BANDPLOT statement 211 LINECHART statement 529

BARCHART statement 232 NEEDLEPLOT statement 593
BARCHARTPARM statement 267 SCATTERPLOT statement 695
BOXPLOT statement 317 SERIESPLOT statement 759
BOXPLOTPARM statement 350 STEPPLOT statement 792
BUBBLEPLOT statement 377 GUTTER= option
ELLIPSE statement 426 AXISTABLE statement 194
ELLIPSEPARM statement 434 INNERMARGIN statement 167
FRINGEPLOT statement 441 LAYOUT GLOBALLEGEND
HIGHLOWPLOT statement 480 statement 97
HISTOGRAM statement 408, 500 MOSAICPLOTPARM statement 577
LINECHART statement 529
LINEPARM statement 549
LOESSPLOT statement 559 H
NEEDLEPLOT statement 592 HALIGN= option
PBSPLINEPLOT statement 606 AXISLEGEND statement 1092
PIECHART statement 619 CONTINUOUSLEGEND statement
POLYGONPLOT statement 636 1102
REGRESSIONPLOT statement 673 DISCRETELEGEND statement 1115
SCATTERPLOT statement 694 ENTRYFOOTNOTE statement 1159
SCATTERPLOTMATRIX statement ENTRYTITLE statement 1167
727 LAYOUT GLOBALLEGEND
SERIESPLOT statement 758 statement 98
STEPPLOT statement 791 LAYOUT GRIDDED statement 105
TEXTPLOT statement 819 LAYOUT LATTICE statement 118
VECTORPLOT statement 846 MERGEDLEGEND statement 1115
GROUP100= option HALIGN= prefix option
BARCHART statement 233 ENTRY statement 1138, 1152
BARCHARTPARM statement 268 ENTRYFOOTNOTE statement 1138
GROUPDISPLAY= option ENTRYTITLE statement 1138
BARCHART statement 234 example uses 1138
BARCHARTPARM statement 270 HEADERBACKGROUNDCOLOR=
BOXPLOT statement 318 option
BOXPLOTPARM statement 351 LAYOUT DATALATTICE statement
HIGHLOWPLOT statement 481 52
LINECHART statement 529 LAYOUT DATAPANEL statement 78
NEEDLEPLOT statement 593 HEADERBORDER= option
SCATTERPLOT statement 695 LAYOUT DATALATTICE statement
SERIESPLOT statement 758 52
STEPPLOT statement 791 LAYOUT DATAPANEL statement 74
TEXTPLOT statement 820 HEADERLABEL= option
grouped data AXISTABLE statement 195
cycling through group attributes 183 HEADERLABELATTRS= option
lines, colors, and marker symbols 179 AXISTABLE statement 195
remapping colors, fills, markers, lines LAYOUT DATALATTICE statement
183 53
GROUPGAP= option LAYOUT DATAPANEL statement 78
PIECHART statement 620 HEADERLABELDISPLAY= option
GROUPLABELOPTS= option LAYOUT DATALATTICE statement
PIECHART statement 621 53
GROUPORDER= option LAYOUT DATAPANEL statement 78
BARCHART statement 235 HEADERLABELLOCATION= option
BARCHARTPARM statement 271 LAYOUT DATALATTICE statement
BOXPLOT statement 319 53
BOXPLOTPARM statement 352 HEADEROPAQUE= option
HIGHLOWPLOT statement 482
Index 1393

LAYOUT DATALATTICE statement IMAGE= argument

53 SYMBOLIMAGE statement 1180
LAYOUT DATAPANEL statement 78 images
HEADERPACK= option image format 176
LAYOUT DATALATTICE statement image name 176
54, 79 image size 176
HEADERSEPARATOR= option INCLUDEMISSINGCLASS= option
LAYOUT DATALATTICE statement AXISTABLE statement 195
54, 79 BLOCKPLOT statement 294
HEADERSPLITCOUNT= option LAYOUT DATALATTICE statement
LAYOUT DATALATTICE statement 55
54, 79 LAYOUT DATAPANEL statement 80
HEATMAP statement INCLUDEMISSINGCOLOR= option
TEMPLATE procedure 446 HEATMAPPARM statement 464
HEATMAPPARM statement INCLUDEMISSINGDISCRETE= option
TEMPLATE procedure 459 BEGINGRAPH statement 30
height, specifying on a graph 16, 176 INCLUDEMISSINGGROUP= option
HEIGHT= argument BANDPLOT statement 212
DRAWOVAL statement 1235 BARCHART statement 236
DRAWRECTANGLE statement 1243 BARCHARTPARM statement 271
HEIGHT= option BOXPLOT statement 319
DRAWIMAGE statement 1222 BOXPLOTPARM statement 353
ODS GRAPHICS statement 16 BUBBLEPLOT statement 378
HEIGHT= option, ODS GRAPHICS DENSITYPLOT statement 409, 502
statement 176 ELLIPSE statement 427
HEIGHTUNIT= option ELLIPSEPARM statement 435
DRAWIMAGE statement 1222 FRINGEPLOT statement 441
DRAWOVAL statement 1237 HIGHLOWPLOT statement 484
DRAWRECTANGLE statement 1245 LINECHART statement 530
HIGH= argument LINEPARM statement 550
HIGHLOWPLOT statement 474 LOESSPLOT statement 560
HIGHCAP= option NEEDLEPLOT statement 594
HIGHLOWPLOT statement 483 PBSPLINEPLOT statement 607
HIGHLABEL= option PIECHART statement 621
HIGHLOWPLOT statement 484 POLYGONPLOT statement 636
HIGHLOWPLOT statement REGRESSIONPLOT statement 673
TEMPLATE procedure 471 SCATTERPLOT statement 696
HISTOGRAM statement SCATTERPLOTMATRIX statement
TEMPLATE procedure 493 727
HISTOGRAMPARM statement SERIESPLOT statement 760
TEMPLATE procedure 506 STEPPLOT statement 792
HOFFSET= option TEXTPLOT statement 820
SYMBOLCHAR statement 1174 VECTORPLOT statement 847
SYMBOLIMAGE statement 1180 INCLUDERANGES=
LINEAROPTS= option, overlay axis
913
I TIMEOPTS= option, overlay axis 935
ID= argument INDENT= option
POLYGONPLOT statement 631 AXISTABLE statement 196
IF statement INDENTWEIGHT= option
conditional logic 1333 AXISTABLE statement 196
conditional logic on templates 14 INDEX= option
IFC function 1322 BANDPLOT statement 212
IFN function 1322 BARCHART statement 236
IGNORECASE= option BARCHARTPARM statement 272
DISCRETEATTRMAP statement 1288 BOXPLOT statement 320
1394 Index

BOXPLOTPARM statement 353 CONTINUOUSLEGEND statement

BUBBLEPLOT statement 378 1102
ELLIPSEPARM statement 435 INTERPOLATION= regression option
FRINGEPLOT statement 442 LOESSPLOT statement 563
HIGHLOWPLOT statement 484 INTERVAL=
LINECHART statement 530 TIMEOPTS= option, datalattice or
LINEPARM statement 550 datapanel axis 1075
LOESSPLOT statement 560 TIMEOPTS= option, lattice axis 1002
NEEDLEPLOT statement 594 TIMEOPTS= option, overlay axis 937
PBSPLINEPLOT statement 607 TIMEOPTS= option, overlay3d axis
REGRESSIONPLOT statement 674 958
remapping colors, fills, markers, lines INTERVALBARWIDTH= option
183 BARCHART statement 237
SCATTERPLOT statement 696 BARCHARTPARM statement 272
SCATTERPLOTMATRIX statement HIGHLOWPLOT statement 485
728 INTERVALBOXWIDTH= option
SERIESPLOT statement 760 BOXPLOT statement 320
STEPPLOT statement 793 BOXPLOTPARM statement 354
TEXTPLOT statement 820 INTERVALMULTIPLIER=
VECTORPLOT statement 847 TIMEOPTS= option, datalattice or
INITIALBARATTRS= option datapanel axis 1077
WATERFALLCHART statement 865 TIMEOPTS= option, lattice axis 1003
INITIALBARTICKVALUE= option TIMEOPTS= option, overlay axis 938
WATERFALLCHART statement 866 TIMEOPTS= option, overlay3d axis
INITIALBARVALUE= option 959
WATERFALLCHART statement 866 ITEMSIZE= option
INNERMARGIN statement DISCRETELEGEND statement 1115
TEMPLATE procedure 166
INSET= option
LAYOUT DATALATTICE statement J
55 java environment
LAYOUT DATAPANEL statement 80 See memory management
SCATTERPLOTMATRIX statement JITTER= option
728 SCATTERPLOT statement 697
INSETOPTS= option JITTEROPTS= option
LAYOUT DATALATTICE statement SCATTERPLOT statement 698
56 JOIN= option
LAYOUT DATAPANEL statement 81 STEPPLOT statement 793
SCATTERPLOTMATRIX statement JREOPTIONS option
729 managing Java memory 1363
INSIDEVALUEATTRS= option JUSTIFY= option
MOSAICPLOTPARM statement 578 BANDPLOT statement 212
integer value type 1341 DRAWTEXT statement 1252
integer-column value type 1341 STEPPLOT statement 793
INTEGER=
COMMONAXISOPTS= option,
overlayequated axis 1014 K
LINEAROPTS= option, datalattice or KERNEL distribution option
datapanel axis 1056 DENSITYPLOT statement 412
LINEAROPTS= option, lattice axis 984 KURTOSIS function 1328
LINEAROPTS= option, overlay axis
915
LINEAROPTS= option, overlay3d axis L
951 LABEL=
INTEGER= option XAXISOPTS=, YAXISOPTS= options,
overlayequated axis 1022
Index 1395

LABEL= option POLYGONPLOT statement 640

AXISTABLE statement 197 LABELSPLITCHARDROP= option
BLOCKPLOT statement 295 datalattice, datapanel axis options 1042
datalattice, datapanel axis options 1038 lattice axis options 971
DROPLINE statement 419 overlay axis options 897
lattice axis options 968 POLYGONPLOT statement 640
LEGENDITEM statement 1128 LABELSPLITJUSTIFY= option
LEGENDTEXTITEMS statement 1132 datalattice, datapanel axis options 1043
overlay axis options 894 lattice axis options 971
overlay3d axis options 948 overlay axis options 898
POLYGONPLOT statement 637 POLYGONPLOT statement 641
LABELATTRS= LABELSTRIP= option
XAXISOPTS=, YAXISOPTS= options, SCATTERPLOT statement 699
overlayequated axis 1023 SCATTERPLOTMATRIX statement
LABELATTRS= option 731
AXISTABLE statement 197 LATTICE layout 43, 111
BLOCKPLOT statement 295 LAYER= option
datalattice, datapanel axis options 1039 BEGINPOLYGON statement 1202
DROPLINE statement 420 BEGINPOLYLINE statement 1208
HIGHLOWPLOT statement 485 DRAWARROW statement 1216
lattice axis options 968 DRAWIMAGE statement 1223
LEGENDITEM statement 1128 DRAWLINE statement 1230
LEGENDTEXTITEMS statement 1132 DRAWOVAL statement 1237
MOSAICPLOTPARM statement 578 DRAWRECTANGLE statement 1245
overlay axis options 894 DRAWTEXT statement 1253
overlay3d axis options 948 LAYOUT DATALATTICE statement 43
POLYGONPLOT statement 637 axis options 1032
LABELFAR= option TEMPLATE procedure 45
BOXPLOT statement 320 LAYOUT DATAPANEL statement 43
BOXPLOTPARM statement 354 axis options 1032
LABELFITPOLICY= option TEMPLATE procedure 70
datalattice, datapanel axis options 1039 LAYOUT GLOBALLEGEND statement
overlay axis options 894, 969 TEMPLATE procedure 96
PIECHART statement 622 LAYOUT GRIDDED statement 43
LABELHALIGN= option TEMPLATE procedure 102
AXISTABLE statement 197 LAYOUT LATTICE statement 43
LABELJUSTIFY= option axis options 963
AXISTABLE statement 198 TEMPLATE procedure 111
LABELLOCATION= option LAYOUT OVERLAY statement 42
POLYGONPLOT statement 638 axis options 889
LABELPLACEMENT= option TEMPLATE procedure 136
BEGINGRAPH statement 30 LAYOUT OVERLAY3D statement 42
LABELPOSITION= option axis options 945
AXISTABLE statement 198 TEMPLATE procedure 152
BLOCKPLOT statement 296 LAYOUT OVERLAYEQUATED
lattice axis options 969 statement 42
overlay axis options 895, 1040 axis options 1013
POLYGONPLOT statement 638 TEMPLATE procedure 144
labels LAYOUT PROTOTYPE statement 42
location and position in templates 185 TEMPLATE procedure 159
LABELSPLIT= option with LAYOUT DATALATTICE 68
POLYGONPLOT statement 639 with LAYOUT DATAPANEL 93
LABELSPLITCHAR= option LAYOUT REGION statement 42
datalattice, datapanel axis options 1041 TEMPLATE procedure 162
lattice axis options 970 layout statements
overlay axis options 896 conditional logic 14, 1333
1396 Index

multi-cell 8 LIMITLOWER= argument

overlay-type 8 BANDPLOT statement 206
summary 41 LIMITUPPER= argument
LCLM function 1328 BANDPLOT statement 206
LEGENDITEM statement line
TEMPLATE procedure 1126 drawing a line 1227
LEGENDLABEL= option line attributes
BANDPLOT statement 213 available line patterns 1352
BARCHART statement 237 specifying line color 1349
BARCHARTPARM statement 272 specifying line pattern 177, 1349
BIHISTOGRAM3DPARM statement specifying line thickness 1349
286 with grouped data 179
BOXPLOT statement 321 with non-grouped data 177
BOXPLOTPARM statement 354 line styles
BUBBLEPLOT statement 378 remapping in grouped data 183
CONTOURPLOTPARM statement 390 line-pattern-name value type 1341
DENDROGRAM statement 397 line-pattern-number value type 1341
DENSITYPLOT statement 410, 502 LINEAROPTS= option
DROPLINE statement 420 datalattice, datapanel axis options 1043
ELLIPSE statement 427 lattice axis options 971
ELLIPSEPARM statement 436 overlay axis options 898
FRINGEPLOT statement 442 overlay3d axis options 948
HISTOGRAM statement 502 LINEATTRS= option
HISTOGRAMPARM statement 516 BEGINPOLYLINE statement 1209
LINECHART statement 531 CONTOURPLOTPARM statement 391
LINEPARM statement 550 DENDROGRAM statement 398
LOESSPLOT statement 561 DENSITYPLOT statement 410
MODELBAND statement 570 DRAWARROW statement 1216
MOSAICPLOTPARM statement 579 DRAWLINE statement 1230
NEEDLEPLOT statement 595 DROPLINE statement 420
PBSPLINEPLOT statement 608 FRINGEPLOT statement 442
POLYGONPLOT statement 641 HIGHLOWPLOT statement 485
REFERENCELINE statement 662 LEGENDITEM statement 1128
REGRESSIONPLOT statement 674 LINECHART statement 531
SCATTERPLOT statement 700 LINEPARM statement 551
SERIESPLOT statement 760 LOESSPLOT statement 561
STEPPLOT statement 794 NEEDLEPLOT statement 595
SURFACEPLOTPARM statement 806 PBSPLINEPLOT statement 608
TEXTPLOT statement 821 REFERENCELINE statement 662
VECTORPLOT statement 848 REGRESSIONPLOT statement 674
WATERFALLCHART statement 867 SERIESPLOT statement 761
legends STEPPLOT statement 794
AXISLEGEND statement 1089 VALUE statement 1290
CONTINUOUSLEGEND statement VECTORPLOT statement 848
1098 LINECHART statement
defined 8 TEMPLATE procedure 521
DISCRETELEGEND statement 1109 LINECOLORGROUP= option
MERGEDLEGEND statement 1109 SERIESPLOT statement 761
overview 11 LINEEXTENT= option
LEGENDTEXTITEMS statement COMMONAXISOPTS= option,
TEMPLATE procedure 1131 overlayequated axis 1014
LEGENDTITLEPOSITION= option overlay axis options 898
LAYOUT GLOBALLEGEND XAXISOPTS=, YAXISOPTS= options,
statement 98 overlayequated axis 1023
LEVELS= option LINELABELATTRS= option
CONTOURPLOTPARM statement 391 CONTOURPLOTPARM statement 391
Index 1397

LINELABELBASELINE= option specifying marker color 1350

CONTOURPLOTPARM statement 391 specifying marker size 1350
LINELABELFORMAT= option specifying marker symbol 1350
CONTOURPLOTPARM statement 392 specifying marker weight 1351
LINELABELPOSITION= option with grouped data 179
CONTOURPLOTPARM statement 392 with non-grouped data 177
LINEPARM statement marker-name value type 1341
TEMPLATE procedure 542 MARKERATTRS= option
LINEPATTERNGROUP= option LEGENDITEM statement 1128
SERIESPLOT statement 762 LINECHART statement 532
LINETHICKNESSMAX= option NEEDLEPLOT statement 595
SERIESPLOT statement 763 SCATTERPLOT statement 700
STEPPLOT statement 794 SCATTERPLOTMATRIX statement
VECTORPLOT statement 848 731
LINETHICKNESSMAXRESPONSE= SERIESPLOT statement 764
option STEPPLOT statement 796
SERIESPLOT statement 763 VALUE statement 1290
STEPPLOT statement 795 MARKERCHARACTER= option
VECTORPLOT statement 849 SCATTERPLOT statement 700
LINETHICKNESSMIN= option SCATTERPLOTMATRIX statement
SERIESPLOT statement 763 732
STEPPLOT statement 795 MARKERCHARACTERATTRS= option
VECTORPLOT statement 849 SCATTERPLOT statement 701
LINETHICKNESSRESPONSE= option SCATTERPLOTMATRIX statement
SERIESPLOT statement 764 732
STEPPLOT statement 796 MARKERCHARACTERPOSITION=
VECTORPLOT statement 850 option
LOCATION= option SCATTERPLOT statement 702, 733
AXISLEGEND statement 1093 MARKERCOLORGRADIENT= option
CONTINUOUSLEGEND statement SCATTERPLOT statement 702
1103 SCATTERPLOTMATRIX statement
DISCRETELEGEND statement 1117 733
LAYOUT GRIDDED statement 105 MARKERCOLORGROUP= option
MERGEDLEGEND statement 1117 SERIESPLOT statement 765
LOESSPLOT statement MARKERFILLATTRS= option
TEMPLATE procedure 553 LINECHART statement 532
LOGOPTS= option SCATTERPLOT statement 703
datalattice, datapanel axis options 1043 SERIESPLOT statement 766
lattice axis options 971 STEPPLOT statement 797
overlay axis options 899 MARKEROUTLINEATTRS= option
LOW= argument LINECHART statement 533
HIGHLOWPLOT statement 474 SCATTERPLOT statement 703
LOWCAP= option SERIESPLOT statement 766
HIGHLOWPLOT statement 485 STEPPLOT statement 797
LOWLABEL= option MARKERSIZEMAX= option
HIGHLOWPLOT statement 486 SCATTERPLOT statement 703
MARKERSIZEMIN= option
SCATTERPLOT statement 704
M MARKERSIZERESPONSE= option
macro variables SCATTERPLOT statement 704
compared to dynamic variables 1315 MARKERSYMBOLGROUP= option
overview 13 SERIESPLOT statement 767
using on templates 1313 MATRIXTYPE= option
marker attributes SCATTERPLOTMATRIX statement
remapping in grouped data 183 734
specifying in a template 178 MAX function 1328
1398 Index

MAXPOINTS= regression option XAXISOPTS=, YAXISOPTS= options,

LOESSPLOT statement 563 overlayequated axis 1024
PBSPLINEPLOT statement 611 MINORTICKCOUNT=
REGRESSIONPLOT statement 677 LINEAROPTS= option, datalattice or
MEAN function 1328 datapanel axis 1058
MEANATTRS= option LINEAROPTS= option, lattice axis 985
BOXPLOT statement 321 LINEAROPTS= option, overlay axis
BOXPLOTPARM statement 355 916
MEDIAN function 1328 LINEAROPTS= option, overlay3d axis
MEDIANATTRS= option 952
BOXPLOT statement 321 LOGOPTS= option, datalattice or
BOXPLOTPARM statement 355 datapanel axis 1068
memory management LOGOPTS= option, lattice axis 995
in java environment 1363 LOGOPTS= option, overlay axis 927
JREOPTIONS option 1363 XAXISOPTS=, YAXISOPTS= options,
MERGEDLEGEND statement overlayequated axis 1025
TEMPLATE procedure 1109 MINORTICKINTERVAL=
MIN function 1328 TIMEOPTS= option, lattice axis 1005
MINORGRID= TIMEOPTS= option, overlay axis 940
LINEAROPTS= option, datalattice or TIMEOPTS= option, overlay3d axis
datapanel axis 1056 961
LINEAROPTS= option, lattice axis 984 TTIMEOPTS= option, datalattice or
LINEAROPTS= option, overlay axis datapanel axis 1078
915 MINORTICKS=
LINEAROPTS= option, overlay3d axis LINEAROPTS= option, datalattice or
951 datapanel axis 1058
LOGOPTS= option, datalattice or LINEAROPTS= option, lattice axis 986
datapanel axis 1066 LINEAROPTS= option, overlay axis
LOGOPTS= option, lattice axis 994 917
LOGOPTS= option, overlay axis 926 LINEAROPTS= option, overlay3d axis
TIMEOPTS= option, datalattice or 953
datapanel axis 1077 LOGOPTS= option, datalattice or
TIMEOPTS= option, lattice axis 1004 datapanel axis 1068
TIMEOPTS= option, overlay axis 938 LOGOPTS= option, lattice axis 996
TIMEOPTS= option, overlay3d axis LOGOPTS= option, overlay axis 928
960 TIMEOPTS= option, datalattice or
XAXISOPTS=, YAXISOPTS= options, datapanel axis 1078
overlayequated axis 1024 TIMEOPTS= option, lattice axis 1005
MINORGRIDATTRS= TIMEOPTS= option, overlay axis 940
LINEAROPTS= option, datalattice or TIMEOPTS= option, overlay3d axis
datapanel axis 1057 961
LINEAROPTS= option, lattice axis 985 MINORTICKS= option
LINEAROPTS= option, overlay axis XAXISOPTS=, YAXISOPTS= options,
916 overlayequated axis 1025
LINEAROPTS= option, overlay3d axis MODELBAND statement
952 TEMPLATE procedure 565
LOGOPTS= option, datalattice or MODELNAME= option
datapanel axis 1067 BANDPLOT statement 213
LOGOPTS= option, lattice axis 995 MOSAICPLOTPARM statement
LOGOPTS= option, overlay axis 927 TEMPLATE procedure 573
TIMEOPTS= option, datalattice or multi-cell layouts 8
datapanel axis 1077 MVAR statement
TIMEOPTS= option, lattice axis 1004 TEMPLATE procedure 13, 1313
TIMEOPTS= option, overlay axis 939
TIMEOPTS= option, overlay3d axis
960
Index 1399

N NEEDLEPLOT statement
N function 1328 TEMPLATE procedure 584
NAME= argument NHINT= option
DISCRETEATTRMAP statement 1288 CONTOURPLOTPARM statement 392
LEGENDITEM statement 1127 NKNOTS= regression option
LEGENDTEXTITEMS statement 1131 PBSPLINEPLOT statement 611
RANGEATTRMAP statement 1301 NLEVELS= option
SYMBOLCHAR statement 1173 CONTOURPLOTPARM statement 392
SYMBOLIMAGE statement 1180 NMISS function 1328
NAME= option NMVAR statement
AXISTABLE statement 198 TEMPLATE procedure 13, 1313
BANDPLOT statement 213 NODEID= argument
BARCHART statement 237 DENDROGRAM statement 396
BARCHARTPARM statement 273 non-grouped data
BIHISTOGRAM3DPARM statement lines, colors, and marker symbols 177
287 NORMAL distribution option
BLOCKPLOT statement 296 DENSITYPLOT statement 413
BOXPLOT statement 322 NOTES statement
BOXPLOTPARM statement 355 TEMPLATE procedure 13, 1313
BUBBLEPLOT statement 379 number value type 1341
CONTOURPLOTPARM statement 392 NUMERATE function 1325
datalattice, datapanel axis options 1043 numeric-column value type 1341
DENDROGRAM statement 398 NXBINS= option
DENSITYPLOT statement 410 HEATMAP statement 451
DROPLINE statement 420 NYBINS= option
ELLIPSE statement 427 HEATMAP statement 451
ELLIPSEPARM statement 436
FRINGEPLOT statement 443
HEATMAP statement 451 O
HEATMAPPARM statement 464 ODS Graphics
HIGHLOWPLOT statement 486 managing java memory 1363
HISTOGRAM statement 502 ODS GRAPHICS statement
HISTOGRAMPARM statement 516 BORDER= option 16
lattice axis options 899, 972 DISCRETEMAX= option 396
LINECHART statement 533 HEIGHT= option 16, 176, 706
LINEPARM statement 551 IMAGENAME= option 176
LOESSPLOT statement 561 overview 15, 176
MODELBAND statement 570 sizing a graph 16, 176
MOSAICPLOTPARM statement 579 specifying a graph border 16
NEEDLEPLOT statement 596 WIDTH= option 16, 176
PBSPLINEPLOT statement 608 ODS styles
PIECHART statement 622 customizing a style 180
POLYGONPLOT statement 642 defining a style for box plots 330
REFERENCELINE statement 662 GraphBox style element 330
REGRESSIONPLOT statement 675 GraphBoxMean style element 330
SCATTERPLOT statement 704 GraphBoxMedian style element 330
SCATTERPLOTMATRIX statement GraphBoxOutlier style element 330
735 GraphBoxWhisker style element 331
SERIESPLOT statement 768 overriding style element defaults 1139
STEPPLOT statement 798 use in templates 16
SURFACEPLOTPARM statement 806 OFFSETMAX= option
TEXTPLOT statement 821 datalattice, datapanel axis options 1043
VECTORPLOT statement 850 lattice axis options 972
WATERFALLCHART statement 867 overlay axis options 899
NBINS= option overlay3d axis options 948
HISTOGRAM statement 503
1400 Index

XAXISOPTS=, YAXISOPTS= options, LINEAROPTS= option, overlay axis

overlayequated axis 1026 917
OFFSETMIN= option OTHERSLICE= option
datalattice, datapanel axis options 1044 PIECHART statement 622
lattice axis options 972 OTHERSLICEOPTS= option
overlay axis options 900 PIECHART statement 622
overlay3d axis options 949 OUTERPAD= option
XAXISOPTS=, YAXISOPTS= options, AXISLEGEND statement 1094
overlayequated axis 1026 CONTINUOUSLEGEND statement
OPAQUE= option 1104
and draw statements 1196 DISCRETELEGEND statement 1118
AXISLEGEND statement 1093 ENTRY statement 1150
BEGINGRAPH statement 31 ENTRYFOOTNOTE statement 1157
CONTINUOUSLEGEND statement ENTRYTITLE statement 1164
1103 LAYOUT DATALATTICE statement
DISCRETELEGEND statement 1117 59
ENTRY statement 1150 LAYOUT DATAPANEL statement 84
ENTRYFOOTNOTE statement 1157 LAYOUT GLOBALLEGEND
ENTRYTITLE statement 1164 statement 98
INNERMARGIN statement 167 LAYOUT GRIDDED statement 106
LAYOUT DATALATTICE statement LAYOUT LATTICE statement 119
58 LAYOUT OVERLAY statement 139
LAYOUT DATAPANEL statement 84 LAYOUT OVERLAY3D statement 155
LAYOUT GRIDDED statement 106 LAYOUT OVERLAYEQUATED
LAYOUT LATTICE statement 118 statement 147
LAYOUT OVERLAY statement 139 LAYOUT REGION statement 163
LAYOUT OVERLAY3D statement 155 MERGEDLEGEND statement 1118
LAYOUT OVERLAYEQUATED OUTLIERATTRS= option
statement 147 BOXPLOT statement 322
LAYOUT REGION statement 163 BOXPLOTPARM statement 356
MERGEDLEGEND statement 1117 OUTLIERTIP= option
OPEN= option BOXPLOT statement 322
HIGHLOWPLOT statement 486 BOXPLOTPARM statement 356
options OUTLINEATTRS= option
value types 1339 BANDPLOT statement 213
ORDER= option BARCHART statement 238
AXISLEGEND statement 1093 BARCHARTPARM statement 273
DISCRETELEGEND statement 1118 BEGINPOLYGON statement 1202
LAYOUT DATAPANEL statement 84 BIHISTOGRAM3DPARM statement
LAYOUT GRIDDED statement 106 287
LAYOUT LATTICE statement 118 BLOCKPLOT statement 296
MERGEDLEGEND statement 1118 BOXPLOT statement 323
ORIENT= option BOXPLOTPARM statement 356
BARCHART statement 237 BUBBLEPLOT statement 379
BARCHARTPARM statement 273 DRAWOVAL statement 1237
BOXPLOT statement 322 DRAWRECTANGLE statement 1245
BOXPLOTPARM statement 355 ELLIPSE statement 427
CONTINUOUSLEGEND statement ELLIPSEPARM statement 436
1103 HEATMAP statement 452
DENDROGRAM statement 398 HEATMAPPARM statement 464
DENSITYPLOT statement 410 HIGHLOWPLOT statement 486
HISTOGRAM statement 503 HISTOGRAM statement 503
HISTOGRAMPARM statement 516 HISTOGRAMPARM statement 516
LINECHART statement 533 LEGENDITEM statement 1129
ORIGIN= MODELBAND statement 570
MOSAICPLOTPARM statement 579
Index 1401

PIECHART statement 624 LAYOUT DATAPANEL statement 85

POLYGONPLOT statement 642 parameterized plots 10
TEXTPLOT statement 822 PARENTID= argument
WATERFALLCHART statement 867 DENDROGRAM statement 396
OUTLINEDMARKERCHARACTERS= PATTERN= attribute
option for lines 1349
SCATTERPLOT statement 704 PBSPLINEPLOT statement
output on templates 15 TEMPLATE procedure 600
oval PERCENTILE= option
drawing an oval 1233 BOXPLOT statement 323
OVERLAY layout 42, 136, 144, 166 percentiles
overlay-type layouts calculating 329
cycling through group attributes 183 empirical distribution function 329
overview 8 weighted average 329
OVERLAY3D layout 42, 152 PIECHART statement
OVERLAYEQUATED layout 42 TEMPLATE procedure 613
plot area
and curve label location 185
P plot area, defined 7
P1 function 1328 plot statements
P25 function 1328 2D and 3D 10
P5 function 1328 computed 10
P50 function 1328 conditional logic 14, 1333
P75 function 1328 expressions on 1318
P90 function 1328 overview 10
P95 function 1328 parameterized 10
P99 function 1328 value types for options 1339
PAD= option plot wall
AXISLEGEND statement 1094 and draw statements 1196
AXISTABLE statement 198 plots, defined 8
BEGINGRAPH statement 31 polygon
CONTINUOUSLEGEND statement drawing a polygon 1199
1104 POLYGONPLOT statement
DISCRETELEGEND statement 1118 TEMPLATE procedure 628
DRAWTEXT statement 1253 polyline
ENTRY statement 1150 drawing a polyline 1206
ENTRYFOOTNOTE statement 1158 POSITION= option
ENTRYTITLE statement 1165 AXISTABLE statement 199
INNERMARGIN statement 167 TEXTPLOT statement 823
LAYOUT DATALATTICE statement PRIMARY= option
59 BARCHART statement 238
LAYOUT DATAPANEL statement 85 BARCHARTPARM statement 274
LAYOUT GLOBALLEGEND BIHISTOGRAM3DPARM statement
statement 99 287
LAYOUT GRIDDED statement 107 BOXPLOT statement 323
LAYOUT LATTICE statement 119 BOXPLOTPARM statement 357
LAYOUT OVERLAY statement 139 BUBBLEPLOT statement 379
LAYOUT OVERLAY3D statement 156 CONTOURPLOTPARM statement 392
LAYOUT OVERLAYEQUATED DENDROGRAM statement 398
statement 148 DENSITYPLOT statement 410
LAYOUT REGION statement 164 HEATMAP statement 452
MERGEDLEGEND statement 1118 HEATMAPPARM statement 465
TEXTPLOT statement 822 HIGHLOWPLOT statement 487
PANELNUMBER= option HISTOGRAM statement 503
LAYOUT DATALATTICE statement HISTOGRAMPARM statement 517
60 LINECHART statement 533
1402 Index

LOESSPLOT statement 562 BUBBLEPLOT statement 380

NEEDLEPLOT statement 596 remapping colors, fills, markers, lines
PBSPLINEPLOT statement 609 183
POLYGONPLOT statement 642 REPEATEDVALUES= option
REGRESSIONPLOT statement 675 BLOCKPLOT statement 296
SCATTERPLOT statement 705 RESPONSE= argument
SERIESPLOT statement 768 BARCHART statement 221
STEPPLOT statement 798 BARCHARTPARM statement 253
SURFACEPLOTPARM statement 807 PIECHART statement 626
TEXTPLOT statement 824 WATERFALLCHART statement 857
VECTORPLOT statement 850 RESPONSE= optional argument
PROBT function 1329 LINECHART statement 538
PROC TEMPLATE REVERSE=
See TEMPLATE procedure XAXISOPTS=, YAXISOPTS= options,
programming features, runtime 12, 1313, overlayequated axis 1026
1317, 1321, 1333 REVERSE= option
PROTOTYPE layout 42, 159 datalattice, datapanel axis options 1044
with LAYOUT DATALATTICE 68 lattice axis options 973
with LAYOUT DATAPANEL 93 overlay axis options 900
REVERSECOLORMODEL= option
BUBBLEPLOT statement 381
Q CONTOURPLOTPARM statement 393
Q1 function 1329 HEATMAP statement 452
Q3 function 1329 HEATMAPPARM statement 465
QRANGE function 1329 MOSAICPLOTPARM statement 579
quantiles POLYGONPLOT statement 643
empirical distribution function 329 SCATTERPLOT statement 705
weighted average 329 SCATTERPLOTMATRIX statement
735
SURFACEPLOTPARM statement 807
R TEXTPLOT statement 824
RANGE function 1329 REWEIGHT= regression option
RANGE statement 1306 LOESSPLOT statement 564
RANGEALTCOLOR= option ROLENAME= option
RANGE statement 1303 BANDPLOT statement 214
RANGEALTCOLORMODEL= option BARCHARTPARM statement 274
RANGE statement 1304 BOXPLOTPARM statement 357
RANGEATTRMAP block 1282 BUBBLEPLOT statement 381
RANGEATTRMAP statement FRINGEPLOT statement 443
TEMPLATE statement 1301 HEATMAP statement 452
RANGEATTRVAR statement HEATMAPPARM statement 465
TEMPLATE procedure 1308 HIGHLOWPLOT statement 487
RANGECOLOR= option HISTOGRAMPARM statement 517
RANGE statement 1303 LINECHART statement 534
RANGECOLORMODEL= option MOSAICPLOTPARM statement 579
RANGE statement 1305 NEEDLEPLOT statement 597
rectangle POLYGONPLOT statement 643
drawing a rectangle 1241 SCATTERPLOT statement 706
REFERENCELINE statement SCATTERPLOTMATRIX statement
TEMPLATE procedure 654 735
REGION layout 42, 162 SERIESPLOT statement 768
REGRESSIONPLOT statement STEPPLOT statement 798
TEMPLATE procedure 666 TEXTPLOT statement 824
RELATIVESCALE= option VECTORPLOT statement 851
BUBBLEPLOT statement 380 WATERFALLCHART statement 867
RELATIVESCALETYPE= option ROTATE= option
Index 1403

DRAWIMAGE statement 1223 LAYOUT DATAPANEL statement 88

DRAWOVAL statement 1238 LAYOUT GRIDDED statement 108
DRAWRECTANGLE statement 1246 LAYOUT LATTICE statement 122
DRAWTEXT statement 1253 ROWVAR= argument
ENTRY statement 1151 LAYOUT DATALATTICE statement
LAYOUT OVERLAY3D statement 156 48
POLYGONPLOT statement 643 ROWVARS= option
SYMBOLCHAR statement 1174 SCATTERPLOTMATRIX statement
SYMBOLIMAGE statement 1181 735
TEXTPLOT statement 824 ROWWEIGHT= option
ROTATELABEL= option LAYOUT DATALATTICE statement
POLYGONPLOT statement 643 63
ROW2AXES block LAYOUT DATAPANEL statement 88
in LAYOUT LATTICE 132 ROWWEIGHTS= option
ROW2AXISOPTS= option LAYOUT LATTICE statement 122
LAYOUT DATALATTICE statement runtime programming features 12, 1313,
60 1317, 1321, 1333
LAYOUT DATAPANEL statement 86
ROW2DATARANGE= option
LAYOUT DATALATTICE statement S
61 SAPLACEMENTOPTS= option
LAYOUT DATAPANEL statement 87 BEGINGRAPH statement 32
LAYOUT LATTICE statement 121, SAS formats
130 error message in log 1353
ROW2HEADERS statement unsupported currency formats 1355
LAYOUT LATTICE statement 133 unsupported date and time formats
ROWAXES block 1354
in LAYOUT LATTICE 132 unsupported numeric formats 1353
ROWAXIS statement using 1353
in LAYOUT LATTICE 132 SCALE= option
LAYOUT LATTICE statement 963 DRAWIMAGE statement 1223
ROWAXISOPTS= option HISTOGRAM statement 504
LAYOUT DATALATTICE statement SYMBOLCHAR statement 1175, 1181
60, 1032 VECTORPLOT statement 851
LAYOUT DATAPANEL statement 86, SCATTERPLOT statement
1032 TEMPLATE procedure 679
ROWDATARANGE= option SCATTERPLOTMATRIX statement
LAYOUT DATALATTICE statement TEMPLATE procedure 715
61 SEGMENTLABEL= option
LAYOUT DATAPANEL statement 86 BARCHART statement 238, 275
LAYOUT LATTICE statement 120, SEGMENTLABELATTRS= option
130 BARCHART statement 239
ROWGUTTER= option BARCHARTPARM statement 274
LAYOUT DATALATTICE statement SEGMENTLABELFITPOLICY= option
62 BARCHART statement 240
LAYOUT DATAPANEL statement 87 BARCHARTPARM statement 274
LAYOUT GRIDDED statement 107 SEGMENTLABELFORMAT= option
LAYOUT LATTICE statement 121 BARCHART statement 240
ROWHEADERS statement BARCHARTPARM statement 275
LAYOUT LATTICE statement 133 SEMIMAJOR= argument
ROWHEADERS= option ELLIPSEPARM statement 432
LAYOUT DATALATTICE statement SEMIMINOR= argument
62 ELLIPSEPARM statement 432
ROWS= option SEPARATOR= option
LAYOUT DATALATTICE statement INNERMARGIN statement 168
62 SEPARATORATTRS= option
1404 Index

INNERMARGIN statement 168 SMOOTH= regression option

SERIESPLOT statement LOESSPLOT statement 564
TEMPLATE procedure 740 PBSPLINEPLOT statement 611
SGRENDER procedure 4, 6 SMOOTHCONNECT= option
SHORTLABEL= LINECHART statement 534
XAXISOPTS=, YAXISOPTS= options, SERIESPLOT statement 768
overlayequated axis 1026 SORTBY= option
SHORTLABEL= option DISCRETELEGEND statement 1119
datalattice, datapanel axis options 1044 sorting functions 1325
lattice axis options 973 SORTORDER= option
overlay axis options 900 DISCRETELEGEND statement 1119
SHORTTEXT= option LAYOUT DATALATTICE statement
ENTRYFOOTNOTE statement 1158 64
ENTRYTITLE statement 1166 LAYOUT DATAPANEL statement 90
SHOWMISSING= option MERGEDLEGEND statement 1119
AXISTABLE statement 199 SPACEFILL=
SHRINKFONTS= option SIDEBAR statement, datalattice layout
LAYOUT DATALATTICE statement 66
63 SIDEBAR statement, datapanel layout
LAYOUT DATAPANEL statement 89 93
LAYOUT GRIDDED statement 108 SIDEBAR statement, lattice layout 124
LAYOUT LATTICE statement 123 SPARSE= option
SIDEBAR statement LAYOUT DATAPANEL statement 90
with LAYOUT DATALATTICE 68 special characters
with LAYOUT DATAPANEL 94 escaping 1142
with LAYOUT LATTICE 134 SPLITCHAR= option
SIZE= argument TEXTPLOT statement 826
BUBBLEPLOT statement 370 SPLITCHARDROP= option
SIZE= attribute TEXTPLOT statement 827
for markers 1350 SPLITJUSTIFY= option
for text 1351 TEXTPLOT statement 827
SIZEMAX= option SPLITPOLICY= option
SCATTERPLOT statement 706 TEXTPLOT statement 828
TEXTPLOT statement 825 SPLITTICKVALUE=
SIZEMAXRESPONSE= option TIMEOPTS= option, datalattice or
TEXTPLOT statement 825 datapanel axis 1079
SIZEMIN= option TIMEOPTS= option, lattice axis 1006
SCATTERPLOT statement 706 TIMEOPTS= option, overlay axis 940
TEXTPLOT statement 825 SPLITWIDTH= option
SIZERESPONSE= option TEXTPLOT statement 828
SCATTERPLOT statement 707 SPREAD= option
TEXTPLOT statement 826 BOXPLOT statement 324
SIZETHRESHOLDMAX= option BOXPLOTPARM statement 357
BUBBLEPLOT statement 381 SQUARED= option
SIZEUNIT= option MOSAICPLOTPARM statement 580
DRAWIMAGE statement 1224 START= option
sizing a graph 16, 176 LAYOUT DATALATTICE statement
SKEWNESS function 1329 65
SKIPEMPTYCELLS= option LAYOUT DATAPANEL statement 91
LAYOUT DATALATTICE statement PIECHART statement 624
63 SCATTERPLOTMATRIX statement
LAYOUT DATAPANEL statement 89 735
LAYOUT LATTICE statement 123 STAT= argument
SLOPE= argument BOXPLOTPARM statement 340
ELLIPSEPARM statement 432 STAT= option
LINEPARM statement 544 AXISTABLE statement 200
Index 1405

BARCHART statement 240 SYMBOL= attribute

LINECHART statement 534 for markers 1350
PIECHART statement 624 SYMBOLCHAR statement
WATERFALLCHART statement 868 TEMPLATE procedure 1173
statement arguments SYMBOLIMAGE statement
expressions on 1318 TEMPLATE procedure 1180
statement options
value types 1339
STDDEV function 1329 T
STDERR function 1329 T function 1329
STEPPLOT statement TARGET= option
TEMPLATE procedure 774 BARCHART statement 241
string value type 1341 BARCHARTPARM statement 276
string-column value type 1341 TEMPLATE procedure 4
strings BEGINGRAPH statement 4
special characters in 1142 conditional logic on 14, 1333
STRIP= option DEFINE STATGRAPH statement 4
TEXTPLOT statement 828 DYNAMIC statement 13, 1313
style-reference dynamics on 13, 1313
fill options 1348 expressions 12
line options 1349 flexible templates 12
marker options 1350 functions 12
text options 1351 macro variables on 13, 1313
style-reference value type 1342 MVAR statement 13, 1313
STYLE= attribute NMVAR statement 13, 1313
for text 1351 NOTES statement 13, 1313
styles types of templates 1313
See ODS styles templates
SUB text command anatomy of a graph 7
DRAWTEXT statement 1256 attribute options 177
ENTRY statement 1141, 1152 axes 7, 10
ENTRYFOOTNOTE statement 1141, BEGINGRAPH statement 4
1159 cell 7
ENTRYTITLE statement 1141, 1167 compiling 5
example uses 1141 conditional logic on 14, 1333
SUBPIXEL= option curve labels 185
SCATTERPLOT statement 707, 736 DEFINE STATGRAPH statement 4
subscript text DISPLAY= option 184
See SUB text command dynamics on 13, 1313
SUM function 1329 expressions 12
SUMWGT function 1329 flexible templates 12
SUP text command footnote area 7
DRAWTEXT statement 1256 functions 12
ENTRY statement 1141, 1152 graph 7
ENTRYFOOTNOTE statement 1141, grouped data 179
1159 image format 176
ENTRYTITLE statement 1141, 1167 image name 176
example uses 1141 image size 176
superscript text interactions between options 184
See SUP text command layout statements 8
SURFACECOLORGRADIENT= option legends 8, 11
SURFACEPLOTPARM statement 807 macro variables on 13
SURFACEPLOTPARM statement minimum requirements for a plot 175
TEMPLATE procedure 803 non-grouped data 177
SURFACETYPE= option ODS Graphics environment 176
SURFACEPLOTPARM statement 807 ODS GRAPHICS statement 15
1406 Index

output 15 LINEAROPTS= option, datalattice or

overview 3 datapanel axis 1058
plot area 7 LINEAROPTS= option, lattice axis 986
plot display 177 LINEAROPTS= option, overlay axis
plot types 10 918
plots 8 LINEAROPTS= option, overlay3d axis
rendering a graph 4 953
runtime actions 6 LOGOPTS= option, overlay axis 928
SGRENDER procedure 4, 6 XAXISOPTS=, YAXISOPTS= options,
styles 16 overlayequated axis 1027
title area 7 THRESHOLDMIN=
text LINEAROPTS= option, datalattice or
drawing text 1249 datapanel axis 1059
special characters 1145, 1345 LINEAROPTS= option, lattice axis 986
special characters in 1142 LINEAROPTS= option, overlay axis
text attributes 918
example uses 1139 LINEAROPTS= option, overlay3d axis
specifying text color 1351 954
specifying text font family 1351 LOGOPTS= option, overlay axis 928
specifying text size 1351 XAXISOPTS=, YAXISOPTS= options,
specifying text style 1351 overlayequated axis 1027
specifying text weight 1352 TICKDISPLAYLIST=
text commands DISCRETEOPTS= option, datalattice or
example uses 1140 datapanel axis axis 1048
special characters in 1142 DISCRETEOPTS= option, lattice axis
TEXT= argument 976
LEGENDTEXTITEMS statement 1132 DISCRETEOPTS= option, overlay axis
TEXTPLOT statement 814 905
TEXT= option LINEAROPTS= option, datalattice or
LEGENDITEM statement 1129 datapanel axis 1059
TEXTATTRS= LINEAROPTS= option, lattice axis 987
SYMBOLCHAR statement 1175 LINEAROPTS= option, overlay axis
TEXTATTRS= option 919
LEGENDITEM statement 1129 LINEAROPTS= option, overlay3d axis
LEGENDTEXTITEMS statement 1132 954
TEXTPLOT statement 828 TICKINTERVALSTYLE=
VALUE statement 1291 LOGOPTS= option, datalattice or
TEXTATTRS= prefix option datapanel axis 1068
DRAWTEXT statement 1256 LOGOPTS= option, lattice axis 996
ENTRY statement 1139, 1151, 1152 LOGOPTS= option, overlay axis 929
ENTRYFOOTNOTE statement 1139, TICKSTYLE=
1158, 1159 COMMONAXISOPTS= option,
ENTRYTITLE statement 1139, 1166, overlayequated axis 1015
1167 overlay axis options 901
example uses 1139 TICKTYPE=
TEXTFITPOLICY= option DISCRETEOPTS= option, datalattice or
ENTRYFOOTNOTE statement 1159 datapanel axis axis 1048
ENTRYTITLE statement 1166 DISCRETEOPTS= option, lattice axis
TEXTGROUP= option 976
AXISTABLE statement 200 DISCRETEOPTS= option, overlay axis
TEXTPLOT statement 905
TEMPLATE procedure 811 TICKVALUEATTRS=
THICKNESS= attribute XAXISOPTS=, YAXISOPTS= options,
for lines 1349 overlayequated axis 1028
THRESHOLDMAX= TICKVALUEATTRS= option
datalattice, datapanel axis options 1045
Index 1407

lattice axis options 973 LINEAROPTS= option, datalattice or

overlay axis options 901 datapanel axis 1062
overlay3d axis options 949 LINEAROPTS= option, lattice axis 990
TICKVALUEFITPOLICY= LINEAROPTS= option, overlay axis
DISCRETEOPTS= option, datalattice or 922
datapanel axis 1048 LINEAROPTS= option, overlay3d axis
DISCRETEOPTS= option, lattice axis 956
976 LOGOPTS= option, datalattice or
DISCRETEOPTS= option, overlay axis datapanel axis 1070
906 LOGOPTS= option, lattice axis 998
LINEAROPTS= option, datalattice or LOGOPTS= option, overlay axis 930
datapanel axis 1060 TIMEOPTS= option, datalattice or
LINEAROPTS= option, lattice axis 987 datapanel axis 1081
LINEAROPTS= option, overlay axis TIMEOPTS= option, lattice axis 1007
919 TIMEOPTS= option, overlay axis 942
TIMEOPTS= option, datalattice or TIMEOPTS= option, overlay3d axis
datapanel axis 1079 962
TIMEOPTS= option, lattice axis 1006 TICKVALUEPRIORITY=
TIMEOPTS= option, overlay axis 941 COMMONAXISOPTS= option,
XAXISOPTS=, YAXISOPTS= options, overlayequated axis 1016
overlayequated axis 1028 LINEAROPTS= option, datalattice or
TICKVALUEFORMAT= datapanel axis 1063
DISCRETEOPTS= option, overlay axis LINEAROPTS= option, lattice axis 991
908, 979, 1051 LINEAROPTS= option, overlay axis
LINEAROPTS= option, datalattice or 922
datapanel axis 1060 LINEAROPTS= option, overlay3d axis
LINEAROPTS= option, lattice axis 988 957
LINEAROPTS= option, overlay axis LOGOPTS= option, datalattice or
920 datapanel axis 1071
LINEAROPTS= option, overlay3d axis LOGOPTS= option, lattice axis 998
954 LOGOPTS= option, overlay axis 931
LOGOPTS= option, datalattice or TICKVALUEPRIORITY= option,
datapanel axis 1069 datalattice or datapanel axis 1081
LOGOPTS= option, lattice axis 997 TIMEOPTS= option, lattice axis 1008
LOGOPTS= option, overlay axis 930 TIMEOPTS= option, overlay axis 943
TIMEOPTS= option, datalattice or TICKVALUEROTATION=
datapanel axis 1080 DISCRETEOPTS= option, datalattice or
TIMEOPTS= option, lattice axis 1007 datapanel axis 1052
TIMEOPTS= option, overlay axis 942 DISCRETEOPTS= option, lattice axis
TIMEOPTS= option, overlay3d axis 980
962 DISCRETEOPTS= option, overlay axis
XAXISOPTS=, YAXISOPTS= options, 909
overlayequated axis 1028 LINEAROPTS= option, datalattice or
TICKVALUEHALIGN= option datapanel axis 1063
datalattice, datapanel axis options 1045 LINEAROPTS= option, lattice axis 991
lattice axis options 973 LINEAROPTS= option, overlay axis
overlay axis options 901 923
TICKVALUELIST= TIMEOPTS= option, datalattice or
COMMONAXISOPTS= option, datapanel axis 1082
overlayequated axis 1015 TIMEOPTS= option, overlay axis 943
DISCRETEOPTS= option, datalattice or TTIMEOPTS= option, lattice axis 1008
datapanel axis 1051 TICKVALUESEQUENCE=
DISCRETEOPTS= option, lattice axis COMMONAXISOPTS= option,
979 overlayequated axis 1017
DISCRETEOPTS= option, overlay axis LINEAROPTS= option, datalattice or
908 datapanel axis 1064
1408 Index

LINEAROPTS= option, lattice axis 991 SCATTERPLOTMATRIX statement

LINEAROPTS= option, overlay axis 737
923 SERIESPLOT statement 769
LINEAROPTS= option, overlay3d axis STEPPLOT statement 799
957 TEXTPLOT statement 829
TICKVALUESPLITCHAR= VECTORPLOT statement 851
DISCRETEOPTS= option, datalattice or WATERFALLCHART statement 868
datapanel axis 1052 TIPFORMAT= option
DISCRETEOPTS= option, lattice axis BANDPLOT statement 215
980 BARCHART statement 242
DISCRETEOPTS= option, overlay axis BARCHARTPARM statement 278
909 BOXPLOT statement 325
TICKVALUESPLITCHARDROP= BOXPLOTPARM statement 358
DISCRETEOPTS= option, datalattice or BUBBLEPLOT statement 382
datapanel axis 1053 DENDROGRAM statement 399
DISCRETEOPTS= option, lattice axis DENSITYPLOT statement 411
981 FRINGEPLOT statement 444
DISCRETEOPTS= option, overlay axis HEATMAP statement 453
910 HEATMAPPARM statement 466
TICKVALUESPLITJUSTIFY= HIGHLOWPLOT statement 488
DISCRETEOPTS= option, datalattice or HISTOGRAM statement 504
datapanel axis 1054 HISTOGRAMPARM statement 518
DISCRETEOPTS= option, lattice axis LINECHART statement 535
982 LOESSPLOT statement 562
DISCRETEOPTS= option, overlay axis MODELBAND statement 571
911 MOSAICPLOTPARM statement 580
TICKVALUEVALIGN= option NEEDLEPLOT statement 598
datalattice, datapanel axis options 1045 PBSPLINEPLOT statement 609
lattice axis options 974 PIECHART statement 625
overlay axis options 902 POLYGONPLOT statement 645
TILT= option REGRESSIONPLOT statement 675
LAYOUT OVERLAY3D statement 156 SCATTERPLOT statement 709
TIMEOPTS= option SCATTERPLOTMATRIX statement
datalattice, datapanel axis options 1046 737
lattice axis options 974 SERIESPLOT statement 770
overlay axis options 902 STEPPLOT statement 799
overlay3d axis options 949 TEXTPLOT statement 830
TIP= option VECTORPLOT statement 852
BANDPLOT statement 214 WATERFALLCHART statement 868
BARCHART statement 242 TIPLABEL= option
BARCHARTPARM statement 277 BANDPLOT statement 215
BOXPLOT statement 324 BARCHART statement 243
BOXPLOTPARM statement 358 BARCHARTPARM statement 278
BUBBLEPLOT statement 382 BOXPLOT statement 325
DENDROGRAM statement 398 BOXPLOTPARM statement 359
FRINGEPLOT statement 443 DENDROGRAM statement 399
HEATMAP statement 453 DENSITYPLOT statement 411
HEATMAPPARM statement 466 FRINGEPLOT statement 444
HIGHLOWPLOT statement 487 HEATMAP statement 453
HISTOGRAMPARM statement 517 HEATMAPPARM statement 467
LINECHART statement 534 HIGHLOWPLOT statement 488
MOSAICPLOTPARM statement 580 HISTOGRAM statement 504
NEEDLEPLOT statement 597 HISTOGRAMPARM statement 518
PIECHART statement 625 LINECHART statement 536
POLYGONPLOT statement 644 LOESSPLOT statement 562
SCATTERPLOT statement 708 MODELBAND statement 571
Index 1409

MOSAICPLOTPARM statement 581 in expressions 1319

NEEDLEPLOT statement 598 TYPE= argument
PBSPLINEPLOT statement 609 LEGENDITEM statement 1127
PIECHART statement 626 TYPE= option
POLYGONPLOT statement 645 BANDPLOT statement 215
REGRESSIONPLOT statement 675 datalattice, datapanel axis options 1046
SCATTERPLOT statement 710 DISCRETELEGEND statement 1121
SCATTERPLOTMATRIX statement ELLIPSE statement 428
738 HIGHLOWPLOT statement 489
SERIESPLOT statement 770 lattice axis options 974
STEPPLOT statement 800 LAYOUT GLOBALLEGEND
TEXTPLOT statement 830 statement 100
VECTORPLOT statement 852 overlay axis options 903
WATERFALLCHART statement 869 overlay3d axis options 950
title area, defined 7 TYPEOF function 1326
TITLE= option
AXISLEGEND statement 1095
AXISTABLE statement 200 U
CONTINUOUSLEGEND statement UCLM function 1329
1105 UNICODE text command
DISCRETELEGEND statement 1120 DRAWTEXT statement 1257
LAYOUT GLOBALLEGEND ENTRY statement 1141, 1152
statement 99 ENTRYFOOTNOTE statement 1141,
MERGEDLEGEND statement 1120 1160
TITLEATTRS= option ENTRYTITLE statement 1141, 1167
AXISLEGEND statement 1095 example uses 1141
AXISTABLE statement 200 rule for specification 1142
CONTINUOUSLEGEND statement Unicode values
1105 greek letters, lower case 1143, 1343
DISCRETELEGEND statement 1120 greek letters, upper case 1144, 1344
LAYOUT GLOBALLEGEND hexadecimal values 1142, 1343
statement 99 in user-defined formats 1355
MERGEDLEGEND statement 1120 keywords 1142, 1343
TITLEBORDER= option special characters 1145, 1345
DISCRETELEGEND statement 1120 URL= option
MERGEDLEGEND statement 1120 BARCHART statement 243
TITLEHALIGN= option BARCHARTPARM statement 279
AXISTABLE statement 201 BEGINPOLYGON statement 1203
TITLEJUSTIFY= option BEGINPOLYLINE statement 1209
AXISTABLE statement 201 BOXPLOTPARM statement 359
TRANSPARENCY= attribute BUBBLEPLOT statement 383
for fills 1348 DRAWARROW statement 1216
TRANSPARENCY= option DRAWIMAGE statement 1224
BEGINPOLYGON statement 1203 DRAWLINE statement 1230
BEGINPOLYLINE statement 1209 DRAWOVAL statement 1238
DRAWARROW statement 1216 DRAWRECTANGLE statement 1246
DRAWIMAGE statement 1224 DRAWTEXT statement 1254
DRAWLINE statement 1230 HEATMAPPARM statement 467
DRAWOVAL statement 1238 LINECHART statement 536
DRAWRECTANGLE statement 1246 MOSAICPLOTPARM statement 581
DRAWTEXT statement 1254 NEEDLEPLOT statement 598
TREETYPE= option PIECHART statement 626
DENDROGRAM statement 399 POLYGONPLOT statement 645
TRIMLEADING= option SCATTERPLOT statement 710
DISCRETEATTRMAP statement 1288 SERIESPLOT statement 771
type conversion STEPPLOT statement 800
1410 Index

TEXTPLOT statement 830 BLOCKPLOT statement 297

WATERFALLCHART statement 869 VALUEJUSTIFY= option
USEDISCRETESIZE= option AXISTABLE statement 202
SCATTERPLOT statement 710 VALUELOCATION= option
User-defined formats MOSAICPLOTPARM statement 582
with Unicode values 1355 VALUESTYPE=
USS function 1329 LLOGOPTS= option, lattice axis 999
LOGOPTS= option, datalattice or
datapanel axis 1071
V LOGOPTS= option, overlay axis 932
VALIGN= option VALUEVALIGN= option
AXISLEGEND statement 1095 BLOCKPLOT statement 300
CONTINUOUSLEGEND statement VAR function 1329
1105 VAR= argument
DISCRETELEGEND statement 1121 DISCRETEATTRVAR statement 1297
ENTRY statement 1151 RANGEATTRVAR statement 1309
LAYOUT GRIDDED statement 108 VCENTER= option
LAYOUT LATTICE statement 123 TEXTPLOT statement 831
MERGEDLEGEND statement 1121 VECTORPLOT statement
value types TEMPLATE procedure 837
color 1340 VERTEXLABEL= option
column 1340 LINECHART statement 537
dimension 1340 VERTEXLABELATTRS= option
expression 1341 LINECHART statement 537
for statement options 1339 VERTEXLABELFORMAT= option
format 1341 LINECHART statement 537
integer 1341 VIEWMAX=
integer-column 1341 COMMONAXISOPTS= option,
line-pattern-name 1341 overlayequated axis 1018
line-pattern-number 1341 LINEAROPTS= option, datalattice or
marker-name 1341 datapanel axis 1065
number 1341 LINEAROPTS= option, lattice axis 992
numeric-column 1341 LINEAROPTS= option, overlay axis
string 1341 924
string-column 1341 LOGOPTS= option, datalattice or
style-reference 1342 datapanel axis 1072
VALUE= argument LOGOPTS= option, lattice axis 1000
AXISTABLE statement 192 LOGOPTS= option, overlay axis 932
VALUEATTRS= option TIMEOPTS= option, datalattice or
AXISLEGEND statement 1096 datapanel axis 1082
AXISTABLE statement 201 TIMEOPTS= option, lattice axis 1009
BLOCKPLOT statement 297 TIMEOPTS= option, overlay axis 944
CONTINUOUSLEGEND statement XAXISOPTS=, YAXISOPTS= options,
1106 overlayequated axis 1030
DISCRETELEGEND statement 1122 VIEWMIN=
MERGEDLEGEND statement 1122 COMMONAXISOPTS= option,
MOSAICPLOTPARM statement 581 overlayequated axis 1018
VALUECOUNTHINT= option LINEAROPTS= option, datalattice or
CONTINUOUSLEGEND statement datapanel axis 1065
1106 LINEAROPTS= option, lattice axis 993
VALUEFITPOLICY= option LINEAROPTS= option, overlay axis
BLOCKPLOT statement 297 924
VALUEFORMAT= option LOGOPTS= option, datalattice or
AXISTABLE statement 202 datapanel axis 1073
VALUEHALIGN= option LOGOPTS= option, lattice axis 1000
AXISTABLE statement 202 LOGOPTS= option, overlay axis 933
Index 1411

TIMEOPTS= option, datalattice or ODS GRAPHICS statement 16

datapanel axis 1082 WIDTH= option, ODS GRAPHICS
TIMEOPTS= option, lattice axis 1009 statement 176
TIMEOPTS= option, overlay axis 944 WIDTHUNIT= option
XAXISOPTS=, YAXISOPTS= options, DRAWIMAGE statement 1224
overlayequated axis 1031 DRAWOVAL statement 1239
VOFFSET= option DRAWRECTANGLE statement 1247
SYMBOLCHAR statement 1175 DRAWTEXT statement 1255
SYMBOLIMAGE statement 1181

X
W X= argument
WALLCOLOR= option AXISTABLE statement 192
LAYOUT OVERLAY statement 140 BANDPLOT statement 206
LAYOUT OVERLAY3D statement 156 BARCHART statement 221
LAYOUT OVERLAYEQUATED BARCHARTPARM statement 253
statement 149 BEGINPOLYGON statement 1200
LAYOUT PROTOTYPE statement 161 BEGINPOLYLINE statement 1207
SCATTERPLOTMATRIX statement BIHISTOGRAM3DPARM statement
738 285
WALLDISPLAY= option BLOCKPLOT statement 290
and draw statements 1196 BOXPLOT statement 308
LAYOUT OVERLAY statement 140 BOXPLOTPARM statement 339
LAYOUT OVERLAY3D statement 157 BUBBLEPLOT statement 370
LAYOUT OVERLAYEQUATED CONTOURPLOTPARM statement 388
statement 149 DRAW statement 1200, 1207
LAYOUT PROTOTYPE statement 161 DRAWOVAL statement 1234
SCATTERPLOTMATRIX statement DRAWRECTANGLE statement 1242
738 DROPLINE statement 417
WATERFALLCHART statement ELLIPSE statement 424
TEMPLATE procedure 854 HEATMAP statement 448
WEIGHT= attribute HEATMAPPARM statement 461
for markers 1351 HIGHLOWPLOT statement 474
for text 1352 LINEPARM statement 543
WEIGHT= option LOESSPLOT statement 555
HEATMAP statement 454 NEEDLEPLOT statement 586
HISTOGRAM statement 325, 411, 505 PBSPLINEPLOT statement 602
WEIGHT= regression option POLYGONPLOT statement 631
LOESSPLOT statement 564 REFERENCELINE statement 655
PBSPLINEPLOT statement 611 REGRESSIONPLOT statement 668
REGRESSIONPLOT statement 677 SCATTERPLOT statement 682
WEIGHTS= option SERIESPLOT statement 744
LAYOUT GLOBALLEGEND STEPPLOT statement 778
statement 100 SURFACEPLOTPARM statement 804
WHISKERATTRS= option TEXTPLOT statement 814
BOXPLOT statement 325 VECTORPLOT statement 840
BOXPLOTPARM statement 360 X= option
WHISKERPERCENTILE= option DRAWIMAGE statement 1225
BOXPLOT statement 326 DRAWTEXT statement 1255
width, specifying on a graph 16, 176 X1= argument
WIDTH= argument DRAWARROW statement 1213
DRAWOVAL statement 1235 DRAWLINE statement 1228
DRAWRECTANGLE statement 1243 X1SPACE= option
WIDTH= option DRAWARROW statement 1217
DRAWIMAGE statement 1224 DRAWLINE statement 1231
DRAWTEXT statement 1254 X2= argument
1412 Index

DRAWARROW statement 1213 LAYOUT OVERLAY3D statement

DRAWLINE statement 1229 157, 945
X2AXISOPTS= option LAYOUT OVERLAYEQUATED
LAYOUT OVERLAY statement 141, statement 149, 1013
889 XBINAXIS= option
X2SPACE= option HEATMAP statement 454
DRAWARROW statement 1217 HEATMAPPARM statement 468
DRAWLINE statement 1231 XBINSIZE= option
XAXIS= option HEATMAP statement 455
AXISTABLE statement 202 XBINSTART= option
BANDPLOT statement 216 HEATMAP statement 455
BARCHART statement 244 XBOUNDARIES= option
BARCHARTPARM statement 279 HEATMAPPARM statement 468
BEGINPOLYGON statement 1203, XBOUNDARY= option
1209 HEATMAP statement 455
BLOCKPLOT statement 300 XENDLABELS= option
BOXPLOT statement 326 HEATMAP statement 455
BOXPLOTPARM statement 360 HEATMAPPARM statement 468
BUBBLEPLOT statement 383 XERRORLOWER= option
CONTOURPLOTPARM statement 393 SCATTERPLOT statement 711
DENDROGRAM statement 400 XERRORUPPER= option
DENSITYPLOT statement 412 SCATTERPLOT statement 711
DRAWARROW statement 1216 XGAP= option
DRAWIMAGE statement 1225 HEATMAP statement 455
DRAWLINE statement 1231 HEATMAPPARM statement 468
DRAWOVAL statement 1239 XOFFSET= option
DRAWRECTANGLE statement 1247 POLYGONPLOT statement 646
DRAWTEXT statement 1255 XORIGIN= argument
DROPLINE statement 420 ELLIPSEPARM statement 433
ELLIPSE statement 428 VECTORPLOT statement 840
ELLIPSEPARM statement 436 XSPACE= option
FRINGEPLOT statement 444 BEGINPOLYGON statement 1203,
HEATMAP statement 454 1209
HEATMAPPARM statement 467 DRAWIMAGE statement 1225
HIGHLOWPLOT statement 489 DRAWOVAL statement 1239
HISTOGRAM statement 505 DRAWRECTANGLE statement 1247
HISTOGRAMPARM statement 519 DRAWTEXT statement 1255
LINECHART statement 537 XVALUEFITPOLICY= option
LINEPARM statement 551 MOSAICPLOTPARM statement 582
LOESSPLOT statement 562 XVALUES= option
MODELBAND statement 571 BIHISTOGRAM3DPARM statement
NEEDLEPLOT statement 599 287
PBSPLINEPLOT statement 609 HEATMAP statement 455
POLYGONPLOT statement 646 HEATMAPPARM statement 468
REFERENCELINE statement 662 HISTOGRAM statement 505
REGRESSIONPLOT statement 676 HISTOGRAMPARM statement 519
SCATTERPLOT statement 711
SERIESPLOT statement 771
STEPPLOT statement 801 Y
TEXTPLOT statement 831 Y= argument
VECTORPLOT statement 852 AXISTABLE statement 192
WATERFALLCHART statement 870 BANDPLOT statement 206
XAXISOPTS= option BARCHART statement 221
LAYOUT OVERLAY statement 141, BARCHARTPARM statement 253
889 BEGINPOLYGON statement 1201
BEGINPOLYLINE statement 1207
Index 1413

BIHISTOGRAM3DPARM statement DENDROGRAM statement 400

285 DENSITYPLOT statement 412
BOXPLOT statement 307 DRAWARROW statement 1217
BOXPLOTPARM statement 339 DRAWIMAGE statement 1225
BUBBLEPLOT statement 370 DRAWLINE statement 1231
CONTOURPLOTPARM statement 388 DRAWOVAL statement 1239
DRAW statement 1201, 1207 DRAWRECTANGLE statement 1247
DRAWOVAL statement 1235 DRAWTEXT statement 1256
DRAWRECTANGLE statement 1242 DROPLINE statement 421
DROPLINE statement 417 ELLIPSE statement 428
ELLIPSE statement 424 ELLIPSEPARM statement 437
HEATMAP statement 448 HEATMAP statement 456
HEATMAPPARM statement 461 HEATMAPPARM statement 469
HIGHLOWPLOT statement 474 HIGHLOWPLOT statement 489
LINEPARM statement 543 HISTOGRAM statement 505
LOESSPLOT statement 555 HISTOGRAMPARM statement 519
NEEDLEPLOT statement 587 LINECHART statement 538
PBSPLINEPLOT statement 602 LINEPARM statement 551
POLYGONPLOT statement 631 LOESSPLOT statement 563
REFERENCELINE statement 656 MODELBAND statement 572
REGRESSIONPLOT statement 668 NEEDLEPLOT statement 599
SCATTERPLOT statement 682 PBSPLINEPLOT statement 610
SERIESPLOT statement 744 POLYGONPLOT statement 646
STEPPLOT statement 778 REFERENCELINE statement 663
SURFACEPLOTPARM statement 805 REGRESSIONPLOT statement 676
TEXTPLOT statement 814 SCATTERPLOT statement 711
VECTORPLOT statement 840 SERIESPLOT statement 772
Y= option STEPPLOT statement 801
DRAWIMAGE statement 1225 TEXTPLOT statement 831
DRAWTEXT statement 1255 VECTORPLOT statement 853
Y1= argument WATERFALLCHART statement 870
DRAWARROW statement 1213 YAXISOPTS= option
DRAWLINE statement 1228 LAYOUT OVERLAY statement 141,
Y1SPACE= option 889
DRAWARROW statement 1217 LAYOUT OVERLAY3D statement
DRAWLINE statement 1231 157, 945
Y2= argument LAYOUT OVERLAYEQUATED
DRAWARROW statement 1214 statement 150, 1013
DRAWLINE statement 1229 YBINAXIS= option
Y2AXISOPTS= option HEATMAP statement 456
LAYOUT OVERLAY statement 141, HEATMAPPARM statement 469
889 YBINSIZE= option
Y2SPACE= option HEATMAP statement 456
DRAWARROW statement 1217 YBINSTART= option
DRAWLINE statement 1231 HEATMAP statement 456
YAXIS= option YBOUNDARIES= option
AXISTABLE statement 202 HEATMAPPARM statement 469
BANDPLOT statement 216 YBOUNDARY= option
BARCHART statement 244 HEATMAP statement 456
BARCHARTPARM statement 279 YENDLABELS= option
BEGINPOLYGON statement 1204, HEATMAP statement 457
1210 HEATMAPPARM statement 469
BOXPLOT statement 326 YERRORLOWER= option
BOXPLOTPARM statement 361 SCATTERPLOT statement 712
BUBBLEPLOT statement 384 YERRORUPPER= option
CONTOURPLOTPARM statement 393 SCATTERPLOT statement 712
1414 Index

YGAP= option HEATMAP statement 457

HEATMAP statement 457 HEATMAPPARM statement 470
HEATMAPPARM statement 470
YORIGIN= argument
ELLIPSEPARM statement 433 Z
VECTORPLOT statement 840 Z= argument
YSPACE= option BIHISTOGRAM3DPARM statement
BEGINPOLYGON statement 1204, 285
1210 CONTOURPLOTPARM statement 388
DRAWIMAGE statement 1226 SURFACEPLOTPARM statement 805
DRAWOVAL statement 1239 ZAXISOPTS= option
DRAWRECTANGLE statement 1247 LAYOUT OVERLAY3D statement
DRAWTEXT statement 1256 158, 945
YVALUEFITPOLICY= option ZOOM= option
MOSAICPLOTPARM statement 582 LAYOUT OVERLAY3D statement 158
YVALUES= option
BIHISTOGRAM3DPARM statement
287