How to enable the checking for OUT_OF_ORDER_ITERATION #2302

Jack012a · 2024-12-04T17:37:49Z

Jack012a
Dec 4, 2024

To iterate through an array, you should be at the beginning of the array: to warn you, an OUT_OF_ORDER_ITERATION error is generated when development checks are active.

The simdjson code will set SIMDJSON_DEVELOPMENT_CHECKS=1 in debug mode. Alternatively, you can set the macro SIMDJSON_DEVELOPMENT_CHECKS to 1 prior to including the simdjson.h header to enable these additional checks: just make sure you remove the definition once your code has been tested. When SIMDJSON_DEVELOPMENT_CHECKS is set to 1, the simdjson library runs additional (expensive) tests on your code to help ensure that you are using the library in a safe manner.

#include <iostream>
#define SIMDJSON_DEVELOPMENT_CHECKS 1
#include "simdjson.h"

int main() {
    const char *json = R"({ "people": [ { "name": "Alice", "age": 30 }, { "name": "Bob", "age": 25 } ] })";

    simdjson::ondemand::parser parser;
    simdjson::padded_string jsonbuffer = simdjson::padded_string(json, std::strlen(json));
    auto doc = parser.iterate(jsonbuffer);

    auto person_name2 = doc["people"].at(1)["name"];
    std::cout << person_name2 << std::endl;

    auto person_name = doc["people"].at(0)["name"];
    std::cout << person_name << std::endl;
    return 0;
}

I build this with g++ -std=c++2b -o sibling sibling.cpp simdjson.cpp -g -O0.

Output:
./sibling
"Bob"
"Alice"

Question> I expect to see OUT_OF_ORDER_ITERATION warning but I didn't.
What did I do wrong here?

Thank you

lemire · 2024-12-04T19:32:43Z

lemire
Dec 4, 2024
Maintainer

I expect to see OUT_OF_ORDER_ITERATION warning. What did I do wrong here?

So the documentation you are quoting refers you to the fact that in some cases, you cannot access twice the same array. But in your code, you have two arrays.

It is important to understand that when you do...

auto person_name2 = doc["people"].at(1)["name"];

and later do...

auto person_name = doc["people"].at(0)["name"];

... you are recreating the array (doc["people"]) twice. This is typically a performance anti-pattern. Note that it is generally a performance antipattern, not just in simdjson, but in C++ more broadly. You should not access the same key twice in a sequence: the compiler is unlikely to be able to optimize this away. It is your responsibility as a programmer not to duplicate the work.

Here is a case where OUT_OF_ORDER_ITERATION will be triggered (just one array):

#include <iostream>
#define SIMDJSON_DEVELOPMENT_CHECKS 1
#include "simdjson.h"

int main() {
    const char *json = R"({ "people": [ { "name": "Alice", "age": 30 }, { "name": "Bob", "age": 25 } ] })";

    simdjson::ondemand::parser parser;
     simdjson::padded_string jsonbuffer = simdjson::padded_string(json, std::strlen(json));
    auto doc = parser.iterate(jsonbuffer);
    simdjson::ondemand::array array = doc["people"];
    auto person_name2 = array.at(1)["name"];
    std::cout << person_name2 << std::endl;

    auto person_name = array.at(0)["name"];
    std::cout << person_name << std::endl;
    return 0;
}

In this case, a proper piece of code might look as follows:

#include <iostream>
#include "simdjson.h"

int main() {
    auto json = R"({ "people": [ { "name": "Alice", "age": 30 }, { "name": "Bob", "age": 25 } ] })"_padded;

    simdjson::ondemand::parser parser;
    auto doc = parser.iterate(json);
    for(auto val : doc["people"]) {
      std::cout << val["name"] << std::endl;
    }
    return EXIT_SUCCESS;
}

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

How to enable the checking for OUT_OF_ORDER_ITERATION #2302

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Replies: 1 comment

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Select a reply

Uh oh!

How to enable the checking for OUT_OF_ORDER_ITERATION #2302

Uh oh!

Uh oh!

Jack012a Dec 4, 2024

Replies: 1 comment

Uh oh!

Uh oh!

lemire Dec 4, 2024 Maintainer

Jack012a
Dec 4, 2024

lemire
Dec 4, 2024
Maintainer