HeartSaVioR edited a comment on issue #26162: [SPARK-29438][SS] Use partition ID of StateStoreAwareZipPartitionsRDD for determining partition ID of state store in stream-stream join URL: https://github.com/apache/spark/pull/26162#issuecomment-577067107 All other stateful operations go through StateStoreRDD. https://github.com/apache/spark/blob/f2d71f5838f4a611e3ebd60f49ee80521d772524/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStoreRDD.scala#L57-L75 `StateStoreRDD.getPartitions` is equivalent to `dataRDD.partitions` which dataRDD is most probably ShuffleRDD. Given the list of partitions are being pulled from "child RDD", applying union on streaming aggregation or flatMapGroupsWithState won't break. In below commit I crafted the regression test for streaming aggregation simply. (in different branch) https://github.com/HeartSaVioR/spark/commit/2fd70536e2a92ef01931d551d8a74409a5aaa78f If we would like to prevent such regression for streaming aggregation and flatMapGroupsWithState as well, I'll add UT for flatMapGroupsWithState as well. I feel it may be better to have another PR as this PR fixes the issue whereas new UTs are only preventing regression. WDYT?
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
