Github user gatorsmile commented on a diff in the pull request:
https://github.com/apache/spark/pull/11745#discussion_r56743788
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala
---
@@ -306,14 +311,28 @@ object SetOperationPushDown extends Rule[LogicalPlan]
with PredicateHelper {
}
/**
- * Attempts to eliminate the reading of unneeded columns from the query
plan using the following
- * transformations:
+ * Attempts to eliminate the reading of unneeded columns from the query
plan
+ * by pushing Project through Filter.
*
- * - Inserting Projections beneath the following operators:
- * - Aggregate
- * - Generate
- * - Project <- Join
- * - LeftSemiJoin
+ * Note: This rule could reverse the effects of
PushPredicateThroughProject.
+ * This rule should be run before ColumnPruning for ensuring that Project
can be
+ * pushed as low as possible.
+ */
+object PushProjectThroughFilter extends Rule[LogicalPlan] {
--- End diff --
@davies
The naming of this rule is not right, but I still think this PR fixes the
fundamental issue between the conflicts of `ColumnPruning` and
`PushPredicateThroughProject`. If we do not take the ideas of this PR, I can
find a test case to show the minor fix in `ColumnPruning` does not cover all
the cases.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
