Release Date: 9 July 2019

New features

  • Added procedures to compute memory requirements for UnionFind, LabelPropagation and PageRank as well as for graph loading (algo.memrec)
  • Added different concurrency values for read/write, in addition to compute
  • Added relationship counts and degree distribution statistics to graph load and graph info procedures

Bug fixes

  • PageRank: Fixed null-pointer exception when running on a graph loaded with incoming relationships
  • Huge Graph: Fixed bug in relationship loading, prevented concurrent access to relationship cursors during execution
  • Error handling: when users supply incorrect or non-existent node or relationship labels, algorithms now throw an error instead of executing over the full graph

Improvements

  • Increased histogram accuracy in community detection algorithms
  • Improved node loading for Huge Graphs
  • Changed weight encoding from map to parallel array (faster processing)
  • Louvain algorithm: improved algorithm performance by merging some iterations, reducing allocation, removing some levels of indirection, and using weighted iterators

Recent Graph Data Science Releases

See All Graph Data Science Releases →