Releases: FalkorDB/FalkorDB
v4.10.1
v4.10.0
FalkorDB v4.10.0
FalkorDB v4.10.0 introduces memory-efficient string handling, deeper analytics procedures, and introspection tools to support real-world production scaling and debugging.
Highlights
String Interning
- Added support for string interning to deduplicate identical strings across the database. This reduces memory usage in graphs where string attributes repeat across nodes or edges.
- New scalar function:
intern()
for explicit string interning in queries. - #1095
Use case:
In identity graphs with millions of nodes sharing the same country, email domain, or tags, intern()
can cut memory usage dramatically without changing schema or queries.
Graph Memory Usage Introspection
- New command:
GRAPH.MEMORY USAGE
Returns per-component memory consumption (in MB) for a specified graph—covering nodes, edges, matrices, and indices. - Useful for performance tuning, debugging, and capacity planning.
Use case:
An ops engineer investigating memory pressure in production can now pinpoint whether the bottleneck lies in edge storage, matrix representation, or indexing overhead.
Array & Range Indexing Enhancements
- Improved support for indexing multi-valued fields, including arrays and specialized range types.
- Provides more reliable indexing and query performance across diverse data models.
- Note: This feature is not available in Neo4j.
- #1041
Use case:
When storing documents with multiple classifications (e.g., tags, roles, symptoms), indexing now supports filtering across these array fields with native performance.
New Scalar Function: prev()
- Added
prev()
function for retrieving the previous value in a query sequence—useful in time-series, pagination, and ordering scenarios. - #1113
Use case:
In time-series modeling of user activity, prev()
allows comparison with the previous event in a single query pass—ideal for detecting sudden drops, gaps, or anomalies.
Graph Algorithms
FalkorDB expands its algorithm library to support scalable graph analytics workloads.
CDLP (Community Detection via Label Propagation)
- New procedure:
algo.labelPropagation
- Detects node communities with support for:
- Label and relationship filtering
- Configurable iteration limits
- #1083
Use case:
Identify user clusters in a social network where friendships and shared interactions evolve dynamically—without predefining the number of communities.
Weakly Connected Components (WCC)
- Identifies all connected subgraphs where traversal is possible by ignoring edge direction.
Use case:
In fraud detection, WCC can help isolate disconnected transaction rings or account clusters operating without visible overlap.
Betweenness Centrality
- Measures node influence based on shortest path traversal across the graph.
Use case:
Rank key intermediaries in a supply chain or communication graph to understand bottlenecks or high-leverage actors in the network.
All algorithms support procedure-based invocation.
Miscellaneous Improvements
- Added support for case-insensitive key lookups in map data types.
- UI improvements to the FalkorDB Browser based on community feedback.
Use case:
Developers ingesting mixed-case JSON or external datasets can now perform uniform key access without normalization overhead.
Bug Fixes
- Improved internal handling of edge attribute memory estimation for better efficiency and maintainability. (#1102)
Resources
v4.8.8
v4.8.7
v4.8.6
v4.8.5
v4.8.4
FalkorDB v4.8.4
Core Improvements & Fixes
- Verified that records were properly initialized (#991)
- Refined the algebraic expression used in the conditional traverse reset (#993)
- Optimized index usage to collect only entity attributes (#1005)
- Ensured the effect was consistently replicated when the threshold was zero (#1008)
- Implemented an operation awareness table (#944)
v4.8.3
ABOUT THE VERSION
We've just rolled out FalkorDB version 4.8.3, which brings a set of under-the-hood improvements focused on memory footprint, aggregation performance, and enhanced indexing capabilities.
This update leads to a memory reduction of around 40%. This release is geared towards optimizing resource utilization and query execution.
HIGHLIGHTS
-
We've integrated a new GraphBLAS version, which introduced 32bit matrix indices. V.4.8.3 will be applicable to almost all use cases.
-
Aggregations in general have been improved. For example, the
COLLECT
function has been optimized to reduce the time it takes to collect items in a list. For example, in tests using the query:
UNWIND` range(0, 1000000) AS X RETURN x % 1024, collect({x:x, minus_x: -x})
Version 4.8.3 showed a 65% reduction in processing time.
-
You can now create full-text indexes on edges. This feature enhances search capabilities and supports more complex graph traversals based on textual properties associated with relationships.
-
We've updated FalkorDB-Browser to v1.2.0