[
https://issues.apache.org/jira/browse/HIVE-16208?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Gopal V updated HIVE-16208:
---------------------------
Description: When processing >2x the hash-table size in the vectorized
group-by, the check for fall-back to streaming is wrong because
{{sumBatchSize*minReduction}} is not reset when processing the next split.
(was: When processing >2x the hash-table size in the vectorized group-by, the
check for fall-back to streaming is wrong because {{sumBatchSize*minReduction}}
is not modified by a partial flush or a full flush.)
> Vectorization: ProcessingModeHashAggregate::sumBatchSize is never reset
> -----------------------------------------------------------------------
>
> Key: HIVE-16208
> URL: https://issues.apache.org/jira/browse/HIVE-16208
> Project: Hive
> Issue Type: Bug
> Reporter: Gopal V
>
> When processing >2x the hash-table size in the vectorized group-by, the check
> for fall-back to streaming is wrong because {{sumBatchSize*minReduction}} is
> not reset when processing the next split.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
