Neo4j Release Notes Archive: Graph Data Science

Graph Data Science 2.4.6

15 September 2023

GDS 2.4.6 is compatible with Neo4j 5 & 4.4 versions (≥ 4.4.9). For GDS compatibility with previous releases, please use GDS Compatibility Table.

Ne...

Graph Data Science 2.4.5

24 August 2023

GDS 2.4.5 is compatible with Neo4j 5 & 4.4 versions (≥ 4.4.9). For GDS compatibility with previous releases, please use GDS Compatibility Table.

Bu...

Graph Data Science 2.4.4

16 August 2023

GDS 2.4.4 is compatible with Neo4j 5 & 4.4 versions (≥ 4.4.9). For GDS compatibility with previous releases, please use GDS Compatibility Table.

Bu...

Graph Data Science 2.4.3

26 July 2023

GDS 2.4.3 is compatible with Neo4j 5 & 4.4 versions (≥ 4.4.9). For GDS compatibility with previous releases, please use GDS Compatibility Table.

Im...

Graph Data Science 2.4.1

27 June 2023

Bug fixes

• Fix a bug in K-Core decomposition that can return invalid values if core values are not consecutive.
• Fix a bug when using mutateProper...

Graph Data Science 2.3.8

2 June 2023

GDS 2.3.8 is compatible with Neo4j 5 & 4.4 versions (≥ 4.4.9). For GDS compatibility with previous releases, please use GDS Compatibility Table.

Bug fixes

• Fixes a bug where running Weighted Node Similarity with different relationship types could lead to invalid results.
• Fixes a bug for shortest path algorithms where yielding the `path` parameter for a graph produced by `gds.beta.graph.generate` in stream mode could result in error. The parameter is now returned as null. Affected algorithms include: Dijkstra, AStar, Yens and Delta-Stepping.

...

Graph Data Science 2.3.7

26 May 2023

GDS 2.3.7 is compatible with Neo4j 5 & 4.4 versions (≥ 4.4.9). For GDS compatibility with previous releases, please use GDS Compatibility Table.

Bug fixes

• Fixes a bug in memory estimation for HashGNN when outputDimension is not set. This affects the `gds.beta.hashgnn.mutate.estimate` and `gds.beta.hashgnn.stream.estimate` procedures.
• Fixed a bug which allowed projecting nodes with negative ids through arrow and cypher aggregation and subsequently led to backed-up graphs not being restored.

...

Graph Data Science 2.3.3

21 April 2023

New features

Neo4j Database Compatibility

• This release is compatible with all Neo4j 5.x database version

Graph Data Science 2.3.2

1 January 1970

GDS 2.3.2 is compatible with Neo4j 5 & 4.4 versions (≥ 4.4.9). For GDS compatibility with previous releases, please use GDS Compatibility Table.

New features

Neo4j Database Compatibility

• This release is compatible with all Neo4j 5.x database version <= 5.6.0. Please see our compatibility matrix above.

Bug fixes

• Graphs imported via Arrow no longer cause invalid node mappings that produced ArrayIndexOutOfBoundsExceptions
• Correct memory estimation of Leiden for very small graphs
• KNN no longer result in an AIOOB exception if the array node properties did not exist for some nodes
• CELF no longer returns negative gains for some nodes
• GraphSage will no longer return NaN values because of incorrect neighbor sampling

Improvements

• More accurate memory estimation on Node Similarity and filtered Node Similarity algorithms for high topN or topK values.
• The gds.alpha.modularity procedures for computing modularity no longer require each community ...