Github user dilipbiswal commented on a diff in the pull request:
https://github.com/apache/spark/pull/9844#discussion_r45701500
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala
---
@@ -661,7 +666,9 @@ class Analyzer(
val evaluatedOrderings =
resolvedAliasedOrdering.zip(sortOrder).map {
case (evaluated, order) =>
val index = originalAggExprs.indexWhere {
- case Alias(child, _) => child semanticEquals
evaluated.child
+ case a@Alias(child, _) =>
+ (child semanticEquals evaluated.child) ||
+ a.exprId ==
evaluated.child.asInstanceOf[AttributeReference].exprId
case other => other semanticEquals evaluated.child
--- End diff --
@cloud-fan **THANKS !!** Somehow i kept thinking that we need the full list
of sort attributes for pushdown determination. After your comment , it makes
sense as any resolved sort order attribute(s) must be resolved from the
aggregate expressions and hence its ok to not considered. Right ?
One other question , after our a change a lot of cases are now going
through this codepath and here we add an extra projection above the sort.
Should we add this only when we have added at least one pushdown attribute ? Or
should we fix the tescases instead ?
---
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
