[
https://issues.apache.org/jira/browse/HIVE-4540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Gunther Hagleitner updated HIVE-4540:
-------------------------------------
Attachment: HIVE-4540.1.patch
Problem is that mapjoin.mapreduce optimization leaves un-configured tablescan
in the plan. The metadata only optimization will choke on that. Patch .1 strips
the TS operator out of the plan after merging the child map task into the
current task.
> JOIN-GRP BY-DISTINCT fails with NPE when mapjoin.mapreduce=true
> ---------------------------------------------------------------
>
> Key: HIVE-4540
> URL: https://issues.apache.org/jira/browse/HIVE-4540
> Project: Hive
> Issue Type: Bug
> Reporter: Gunther Hagleitner
> Assignee: Gunther Hagleitner
> Attachments: HIVE-4540.1.patch
>
>
> If the mapjoin.mapreduce optimization kicks in on a query of this form:
> {noformat}
> select count(distinct a.v)
> from a join b on (a.k = b.k)
> group by a.g
> {noformat}
> The planer will NPE in the metadataonly optimizer.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
