java.lang.ArrayIndexOutOfBoundsException while running gds.beta.node2vec.stream

[kinpoon-sn]

Describe the bug

To Reproduce

GDS version: 1.7.2 Neo4j version: 4.3.4 Community Operating system: Amazon Linux rhel fedora

Steps to reproduce the behavior:

• [ ] import data by neo4j-admin import, about 60M nodes, 700M edges
• [ ] query:

  
  CALL gds.beta.node2vec.stream(
        {
          nodeProjection: "*",
          relationshipProjection: "R",
          embeddingDimension: 128,
          randomSeed: 42,
          concurrency: 4,
          walkLength: 10,
          iterations: 1,
          relationshipProperties: ["weight"],
          relationshipWeightProperty: "weight"
        }
      )
      YIELD nodeId, embedding
      RETURN gds.util.asNode(nodeId).nodeId AS name, embedding, labels(gds.util.asNode(nodeId)) AS labels

error log (can run it successfully by fastRP)

2021-11-16 06:48:11.202+0000 INFO [neo4j.BoltWorker-4 [bolt] [/127.0.0.1:43886] ] Node2Vec :: RandomWalk :: create walks 100% 2021-11-16 06:48:11.203+0000 INFO [neo4j.BoltWorker-4 [bolt] [/127.0.0.1:43886] ] Node2Vec :: RandomWalk :: create walks :: Finished 2021-11-16 06:48:11.203+0000 INFO [neo4j.BoltWorker-4 [bolt] [/127.0.0.1:43886] ] Node2Vec :: RandomWalk :: Finished 2021-11-16 06:55:35.822+0000 INFO [neo4j.BoltWorker-4 [bolt] [/127.0.0.1:43886] ] Node2Vec :: train :: Start 2021-11-16 06:55:35.824+0000 INFO [neo4j.BoltWorker-4 [bolt] [/127.0.0.1:43886] ] Node2Vec :: train :: iteration 1 of 1 :: Start 2021-11-16 06:55:42.614+0000 WARN Computation failed java.lang.ArrayIndexOutOfBoundsException: Index 8 out of bounds for length 8 at org.neo4j.gds.core.loading.ZigZagLongDecoding.zigZagUncompress(ZigZagLongDecoding.java:32) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.core.loading.ZigZagLongDecoding.zigZagUncompress(ZigZagLongDecoding.java:25) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.embeddings.node2vec.CompressedRandomWalks$CompressedWalkIterator.next(CompressedRandomWalks.java:141) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.embeddings.node2vec.CompressedRandomWalks$CompressedWalkIterator.next(CompressedRandomWalks.java:95) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.embeddings.node2vec.PositiveSampleProducer.nextWalk(PositiveSampleProducer.java:78) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.embeddings.node2vec.PositiveSampleProducer.nextCenterWord(PositiveSampleProducer.java:100) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.embeddings.node2vec.PositiveSampleProducer.nextContextWord(PositiveSampleProducer.java:122) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.embeddings.node2vec.PositiveSampleProducer.nextContextWord(PositiveSampleProducer.java:125) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.embeddings.node2vec.PositiveSampleProducer.nextContextWord(PositiveSampleProducer.java:125) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.embeddings.node2vec.PositiveSampleProducer.next(PositiveSampleProducer.java:65) ~[neo4j-graph-data-science-1.7.2.jar:?] at org.neo4j.gds.embeddings.node2vec.Node2VecModel$TrainingTask.run(Node2VecModel.java:189) ~[neo4j-graph-data-science-1.7.2.jar:?] at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515) ~[?:?] at java.util.concurrent.FutureTask.run(FutureTask.java:264) ~[?:?] at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128) ~[?:?] at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628) ~[?:?] at java.lang.Thread.run(Thread.java:829) [?:?] at org.neo4j.internal.helpers.NamedThreadFactory$2.run(NamedThreadFactory.java:110) ~[neo4j-common-4.3.4.jar:4.3.4] Suppressed: java.lang.ArrayIndexOutOfBoundsException: Index 8 out of bounds for length 8

[soerenreichardt]

Hello @kinpoon-sn, thanks for the bug report! We were able to locate and fix the issue. The fix should be available in all upcoming patch releases of GDS 1.7 and 1.8.

[kinpoon-sn]

thank you @soerenreichardt , looking forward to the patch