[
https://issues.apache.org/jira/browse/HIVE-21537?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vineet Garg updated HIVE-21537:
-------------------------------
Attachment: HIVE-21537.4.patch
> Scalar query rewrite could be improved to not generate an extra join if
> subquery is guaranteed to produce atmost one row
> ------------------------------------------------------------------------------------------------------------------------
>
> Key: HIVE-21537
> URL: https://issues.apache.org/jira/browse/HIVE-21537
> Project: Hive
> Issue Type: Improvement
> Components: Query Planning
> Affects Versions: 4.0.0
> Reporter: Vineet Garg
> Assignee: Vineet Garg
> Priority: Major
> Labels: sub-query
> Attachments: HIVE-21537.1.patch, HIVE-21537.2.patch,
> HIVE-21537.3.patch, HIVE-21537.4.patch
>
>
> Currently Hive planner introduces this branch and later executes a rule to
> remove this branch if it could.
> Subquery remove rule itself could check if subquery will produce max one row
> (using relmetadat's getMaxRowCount) and avoid introducing this branch.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
