cloud-fan commented on a change in pull request #32602:
URL: https://github.com/apache/spark/pull/32602#discussion_r636702041
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/PropagateEmptyRelation.scala
##########
@@ -72,11 +65,72 @@ object PropagateEmptyRelation extends Rule[LogicalPlan]
with PredicateHelper wit
}
}
+ case p: UnaryNode if p.children.nonEmpty &&
p.children.forall(isEmptyLocalRelation) => p match {
+ case _: Project => empty(p)
+ case _: Filter => empty(p)
+ case _: Sample => empty(p)
+ case _ => p
+ }
+ }
+}
+
+/**
+ * The rule used by both normal Optimizer and AQE Optimizer for:
+ * 1. Binary-node Logical Plans
+ * - Join with one or two empty children (including Intersect/Except).
+ * - Join is single column NULL-aware anti join (NAAJ)
+ * Broadcasted [[HashedRelation]] is [[HashedRelationWithAllNullKeys]].
Eliminate join to an
+ * empty [[LocalRelation]].
+ * - Left semi Join
+ * Right side is non-empty and condition is empty. Eliminate join to its
left side.
+ * - Left anti join
+ * Right side is non-empty and condition is empty. Eliminate join to an
empty
+ * [[LocalRelation]].
+ * 2. Unary-node Logical Plans
+ * - Limit/Repartition with all empty children.
+ * - Aggregate with all empty children and at least one grouping
expression.
+ * - Generate(Explode) with all empty children. Others like Hive UDTF may
return results.
+ *
+ * @param checkRowCount At AQE side, we use the query stage stats to check the
check.
Review comment:
`At AQE side, we use this function to check if a plan has output rows or
not`
--
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]
