karenfeng commented on a change in pull request #31666:
URL: https://github.com/apache/spark/pull/31666#discussion_r608844411
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala
##########
@@ -958,6 +947,34 @@ class Analyzer(override val catalogManager: CatalogManager)
}
}
}
+
+ private def getMetadataAttributes(plan: LogicalPlan): Seq[Attribute] = {
+ lazy val childMetadataOutput = plan.children.flatMap(_.metadataOutput)
+ plan.expressions.flatMap(_.collect {
+ case a: Attribute if a.isMetadataCol => a
+ case a: Attribute if childMetadataOutput.exists(_.exprId == a.exprId)
=>
+ childMetadataOutput.find(_.exprId == a.exprId).get
+ })
+ }
+
+ private def hasMetadataCol(plan: LogicalPlan): Boolean = {
+ lazy val childMetadataOutput = plan.children.flatMap(_.metadataOutput)
+ val hasMetaCol = plan.expressions.exists(_.find {
+ case a: Attribute =>
+ a.isMetadataCol || childMetadataOutput.exists(_.exprId == a.exprId)
Review comment:
This occurs in the case that a column is resolved below the level at
which it becomes labeled as metadata. For the NATURAL/USING JOIN, this happens
when the column is resolved at the level of the root table (this seems to only
happen for DataFrames) - it is only labeled as hidden when it is used as a key
column in the join.
--
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]
