Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
Hive, mail # user - Partition performance


+
Ian 2013-04-04, 23:01
+
Sanjay Subramanian 2013-04-04, 23:06
+
Ramki Palle 2013-04-04, 23:21
+
Owen OMalley 2013-04-04, 23:25
+
Dean Wampler 2013-04-04, 23:28
+
Ian 2013-04-05, 18:36
+
Ramki Palle 2013-04-05, 20:12
+
Ian 2013-04-11, 22:25
+
Peter Marron 2013-07-02, 09:34
+
Owen OMalley 2013-07-02, 14:51
+
David Morel 2013-07-03, 12:19
+
Edward Capriolo 2013-07-03, 14:22
Copy link to this message
-
Re: Partition performance
Owen O'Malley 2013-07-03, 14:56
On Wed, Jul 3, 2013 at 5:19 AM, David Morel <[EMAIL PROTECTED]> wrote:

>
> That is still not really answering the question, which is: why is it slower
> to run a query on a heavily partitioned table than it is on the same number
> of files in a less heavily partitioned table.
>

According to Gopal's investigations in
https://issues.apache.org/jira/browse/HIVE-4051, each time Hive plans a
query, it does a query per a partition to the backing SQL database. That
would explain a lot of the latency for tables with large numbers of
partitions.

-- Owen
+
Peter Marron 2013-07-04, 07:37
+
Peter Marron 2013-07-04, 09:25
+
Dean Wampler 2013-07-03, 13:51