c21 commented on a change in pull request #29277:
URL: https://github.com/apache/spark/pull/29277#discussion_r463395132
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/WholeStageCodegenExec.scala
##########
@@ -903,6 +904,10 @@ case class CollapseCodegenStages(
// The children of SortMergeJoin should do codegen separately.
j.withNewChildren(j.children.map(
child => InputAdapter(insertWholeStageCodegen(child))))
+ case j: ShuffledHashJoinExec =>
+ // The children of ShuffledHashJoin should do codegen separately.
+ j.withNewChildren(j.children.map(
Review comment:
@cloud-fan, @viirya - if we only codegen separately for build side, we
would still have the same problem as above for multiple SHJs right? Essentially
we would fuse multiple stream sides codegen together in one codegen method, so
we will have multiple build side initialized in init(), and naming collision as
above. Let me know if it doesn't make sense, or I can create a counter example
here.
----------------------------------------------------------------
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
