[
https://issues.apache.org/jira/browse/SPARK-45658?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17784567#comment-17784567
]
Asif commented on SPARK-45658:
------------------------------
I also think that during canonicalization of DynamicPruningSubquery, the
pruning key's canonicalization should be done on the basis of the enclosing
Plan which contains the DynamicPruningSubquery Expression
> Canonicalization of DynamicPruningSubquery is broken
> ----------------------------------------------------
>
> Key: SPARK-45658
> URL: https://issues.apache.org/jira/browse/SPARK-45658
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 3.5.0
> Reporter: Asif
> Priority: Major
> Labels: pull-request-available
>
> The canonicalization of (buildKeys: Seq[Expression]) in the class
> DynamicPruningSubquery is broken, as the buildKeys are canonicalized just by
> calling
> buildKeys.map(_.canonicalized)
> The above would result in incorrect canonicalization as it would not be
> normalizing the exprIds relative to buildQuery output
> The fix is to use the buildQuery : LogicalPlan's output to normalize the
> buildKeys expression
> as given below, using the standard approach.
> buildKeys.map(QueryPlan.normalizeExpressions(_, buildQuery.output)),
> Will be filing a PR and bug test for the same.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
