[
https://issues.apache.org/jira/browse/SPARK-60137?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated SPARK-60137:
-----------------------------------
Labels: pull-request-available (was: )
> Allow lossless widening cast in ExprUtils.canEvaluateUnconditionally
> --------------------------------------------------------------------
>
> Key: SPARK-60137
> URL: https://issues.apache.org/jira/browse/SPARK-60137
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 5.0.0
> Reporter: Yuming Wang
> Priority: Major
> Labels: pull-request-available
>
> ExprUtils.canEvaluateUnconditionally checks whether an expression can be
> evaluated unconditionally on additional rows without risking runtime
> failures. Currently, it conservatively excludes all Cast expressions.
> However, lossless numeric widening casts (where
> Cast.canUpCast(child.dataType, toType) is true) are guaranteed not to fail at
> runtime or cause precision loss. We can safely allow such widening casts in
> canEvaluateUnconditionally.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
