[
https://issues.apache.org/jira/browse/HIVE-13788?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rajesh Balamohan updated HIVE-13788:
------------------------------------
Attachment: msck_call_stack_with_fix.png
Thanks for the patch [~hsubramaniyan]. With the patch, it invokes the query in
the DB only once. Attached the profiler output. FileSystem.exists() calls are
expensive as it is contacting S3 in my case.
> hive msck listpartitions need to make use of directSQL instead of datanucleus
> -----------------------------------------------------------------------------
>
> Key: HIVE-13788
> URL: https://issues.apache.org/jira/browse/HIVE-13788
> Project: Hive
> Issue Type: Improvement
> Reporter: Rajesh Balamohan
> Assignee: Hari Sankar Sivarama Subramaniyan
> Priority: Minor
> Attachments: HIVE-13788.1.patch, msck_call_stack_with_fix.png,
> msck_stack_trace.png
>
>
> Currently, for tables having 1000s of partitions too many DB calls are made
> via datanucleus.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)