HyukjinKwon commented on code in PR #41554:
URL: https://github.com/apache/spark/pull/41554#discussion_r1288052991
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/DeduplicateRelations.scala:
##########
@@ -220,7 +220,54 @@ object DeduplicateRelations extends Rule[LogicalPlan] {
if (attrMap.isEmpty) {
planWithNewChildren
} else {
- planWithNewChildren.rewriteAttrs(attrMap)
+ def rewriteAttrsMatchWithSubPlan(
+ attrs: Seq[Attribute],
+ attrMap: AttributeMap[Attribute],
+ planOutput: Seq[Attribute]): Seq[Attribute] = {
+ val canRewriteAttrs = attrMap.filter(a =>
planOutput.contains(a._2))
+ attrs.map(attr => canRewriteAttrs.getOrElse(attr, attr))
+ }
+
+ def rewriteOrderMatchWithSubPlan(
+ attrs: Seq[SortOrder],
+ attrMap: AttributeMap[Attribute],
+ planOutput: Seq[Attribute]): Seq[SortOrder] = {
+ val canRewriteAttrs = attrMap.filter(a =>
planOutput.contains(a._2))
+ attrs.map(attr => {
+
attr.transformWithPruning(_.containsPattern(ATTRIBUTE_REFERENCE)) {
+ case a: AttributeReference =>
+ canRewriteAttrs.getOrElse(a, a)
+ }
+ attr
+ })
+ }
+
+ planWithNewChildren match {
+ case c: CoGroup =>
+ // SPARK-43781: CoGroup is a special case, as it has different
output attributes
Review Comment:
What do you mean by "it has different output attributes"?
There are similar cases where it creates new output attributes (not from the
child): `MapPartitionsInR`, `MapPartitionsInRWithArrow`, `MapElements`,
`MapGroups`, `FlatMapGroupsWithState`, `FlatMapGroupsInR`, `FlatMapGroupsInR`,
`FlatMapGroupsInRWithArrow`, `FlatMapGroupsInPandas`, `MapInPandas`,
`PythonMapInArrow`, and `FlatMapGroupsInPandasWithState`.
There's another CoGroup operation in Python (`cogroup.apply()` that is
`FlatMapCoGroupsInPandas`). Would be great to double check.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
