Github user wzhfy commented on the issue: https://github.com/apache/spark/pull/19855 @gengliangwang @cloud-fan Previously this rule is in the batch `Operator Optimizations`, but after [SPARK-14781](https://github.com/apache/spark/pull/12820), it is moved into a separate batch [by this code change](https://github.com/apache/spark/pull/12820#discussion_r61688622). I think the reason of this change is in the PR description: >This PR also fix a bug in predicate subquery push down through join (they should not). I reverted the change in Optimizer and ran all the test suites in that PR, hoping to find some clue. But I found no error in tests. So it seems something has been changed since that PR.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org