Github user gatorsmile commented on a diff in the pull request: https://github.com/apache/spark/pull/16404#discussion_r93903360 --- Diff: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala --- @@ -1918,28 +1918,37 @@ class Analyzer( case p: Project => p case f: Filter => f + case a: Aggregate if a.groupingExpressions.exists(!_.deterministic) => + val nondeterToAttr = getNondeterToAttr(a.groupingExpressions) + val newChild = Project(a.child.output ++ nondeterToAttr.values, a.child) + a.transformExpressions { case e => + nondeterToAttr.get(e).map(_.toAttribute).getOrElse(e) + }.copy(child = newChild) + // todo: It's hard to write a general rule to pull out nondeterministic expressions // from LogicalPlan, currently we only do it for UnaryNode which has same output // schema with its child. case p: UnaryNode if p.output == p.child.output && p.expressions.exists(!_.deterministic) => --- End diff -- Just a question. What is the reason why we need to have such a condition `p.output == p.child.output`?
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org