Github user aokolnychyi commented on a diff in the pull request:
https://github.com/apache/spark/pull/21473#discussion_r192234621
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala
---
@@ -1744,11 +1744,14 @@ class Analyzer(
* it into the plan tree.
*/
object ExtractWindowExpressions extends Rule[LogicalPlan] {
- private def hasWindowFunction(projectList: Seq[NamedExpression]):
Boolean =
- projectList.exists(hasWindowFunction)
+ private def hasWindowFunction(exprs: Seq[Expression]): Boolean =
+ exprs.exists(hasWindowFunction)
- private def hasWindowFunction(expr: NamedExpression): Boolean = {
+ private def hasWindowFunction(expr: Expression): Boolean = {
expr.find {
+ case AggregateExpression(aggFunc, _, _, _) if
hasWindowFunction(aggFunc.children) =>
--- End diff --
I have some doubts that this is the best place for this check.
StackOverflow happens in ``extract``. We can also define a separate method and
call it inside ``extract``. However, that method will share the same structure
as ``hasWindowFunction``.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
