Re: [PR] [SPARK-46743][SQL] Count bug after constant folding [spark]

["agubichev (via GitHub)" <gi...@apache.org>]

agubichev commented on code in PR #45125:
URL: https://github.com/apache/spark/pull/45125#discussion_r1511830190


##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala:
##########
@@ -328,6 +328,30 @@ abstract class Optimizer(catalogManager: CatalogManager)
       // Do not optimize DPP subquery, as it was created from optimized plan and we should not
       // optimize it again, to save optimization time and avoid breaking broadcast/subquery reuse.
       case d: DynamicPruningSubquery => d
+      case s@ScalarSubquery(a@Aggregate(group, _, child), _, _, _, _, mayHaveCountBug)
+        if mayHaveCountBug.nonEmpty && mayHaveCountBug.get =>
+        // This is a subquery with an aggregate that may suffer from a COUNT bug.
+        // Detailed COUNT bug detection is done at a later stage (e.g. in
+        // RewriteCorrelatedScalarSubquery).
+        // Make sure that the output plan always has the same aggregate node
+        // (i.e., it is not being constant folded).
+        // Note that this does not limit optimization opportunities for the subquery: after
+        // decorrelation is done, the subquery's body becomes part of the main plan and all
+        // optimization rules are applied again.
+        val groupRefs = group.flatMap(x => x.references)
+        val projectOverSubqueryBody = Project(groupRefs ++ a.references.toSeq, child)

Review Comment:
   renamed



-- 
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: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org