tanelk commented on a change in pull request #31980:
URL: https://github.com/apache/spark/pull/31980#discussion_r602735414
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala
##########
@@ -991,13 +991,24 @@ object TransposeWindow extends Rule[LogicalPlan] {
})
}
+ private def windowsCompatible(w1: Window, w2: Window): Boolean = {
+ w1.references.intersect(w2.windowOutputSet).isEmpty &&
+ w1.expressions.forall(_.deterministic) &&
+ w2.expressions.forall(_.deterministic) &&
+ compatiblePartitions(w1.partitionSpec, w2.partitionSpec)
+ }
+
def apply(plan: LogicalPlan): LogicalPlan = plan transformUp {
- case w1 @ Window(we1, ps1, os1, w2 @ Window(we2, ps2, os2, grandChild))
- if w1.references.intersect(w2.windowOutputSet).isEmpty &&
- w1.expressions.forall(_.deterministic) &&
- w2.expressions.forall(_.deterministic) &&
- compatiblePartitions(ps1, ps2) =>
- Project(w1.output, Window(we2, ps2, os2, Window(we1, ps1, os1,
grandChild)))
+ case w1 @ Window(_, _, _, w2 @ Window(_, _, _, grandChild))
+ if windowsCompatible(w1, w2) =>
+ Project(w1.output, w2.copy(child = w1.copy(child = grandChild)))
+
+ case w1 @ Window(_, _, _, Project(pl, w2 @ Window(_, _, _, grandChild)))
+ if windowsCompatible(w1, w2) &&
w1.references.subsetOf(grandChild.outputSet) =>
+ Project(
+ pl ++ w1.windowOutputSet,
Review comment:
The `w1.output` contains only the attributes (name + id), but not how
the attributes are computed based on the input attributes. This information is
in the project list.
--
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]
