[
https://issues.apache.org/jira/browse/HIVE-5158?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13755025#comment-13755025
]
Edward Capriolo commented on HIVE-5158:
---------------------------------------
{quote}
The existing logic to get all partitions doesn't actually appear to have any.
{/quote}
The partitions are ordered so you can start at an existing one and read N
partitions. I am not sure it is related to your issue but I just wanted to
remind everyone that if you doing two levels of partitions the number adds up
very fast and can OOM the client, or the thrift server.
> allow getting all partitions for table to also use direct SQL path
> ------------------------------------------------------------------
>
> Key: HIVE-5158
> URL: https://issues.apache.org/jira/browse/HIVE-5158
> Project: Hive
> Issue Type: Improvement
> Reporter: Sergey Shelukhin
> Assignee: Sergey Shelukhin
> Attachments: HIVE-5158.D12573.1.patch, HIVE-5158.D12573.2.patch,
> HIVE-5158.D12573.3.patch, HIVE-5158.D12573.4.patch
>
>
> While testing some queries I noticed that getPartitions can be very slow
> (which happens e.g. in non-strict mode with no partition column filter); with
> a table with many partitions it can take 10-12s easily. SQL perf path can
> also be used for this path.
--
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
