[
https://issues.apache.org/jira/browse/CALCITE-5174?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18045869#comment-18045869
]
Zhen Chen commented on CALCITE-5174:
------------------------------------
Fixed via CALCITE-7327.
> EnumerableHashJoin filter out NULL values for "IS NOT DISTINCT FROM"
> --------------------------------------------------------------------
>
> Key: CALCITE-5174
> URL: https://issues.apache.org/jira/browse/CALCITE-5174
> Project: Calcite
> Issue Type: Improvement
> Components: core
> Affects Versions: 1.30.0
> Reporter: Vladimir Ozerov
> Priority: Major
>
> The {{Join}} operator constructs the join info from the condition using
> {{JoinInfo.of}}. This method collects information whether {{NULL}} values
> should be filtered for the given pair of left/right key. However, this
> information is not exposed from the {{Join}} operator and also not used by
> the {{EnumerableHashJoin]}, yielding incorrect results for queries like {{ON
> a IS NOT DISTINCT FROM b}}.
> The problem is pretty important, because the {{IS NOT DISTINCT FROM}} may
> silently appear in the realtional tree due to various simplifications and
> optimizations.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
