Mailing-List: contact dev-help@hive.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hive.apache.org
Date: Mon, 4 Feb 2013 01:23:11 +0000 (UTC)
From: "Navis (JIRA)" <jira@apache.org>
To: hive-dev@hadoop.apache.org
Message-ID: <JIRA.12630324.1359704667022.237122.1359940991857@arcas>
In-Reply-To: <JIRA.12630324.1359704667022@arcas>
References: <JIRA.12630324.1359704667022@arcas>
Subject: [jira] [Commented] (HIVE-3972) Support using multiple reducer for
 fetching order by results
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HIVE-3972?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13569944#comment-13569944 ] 

Navis commented on HIVE-3972:
-----------------------------

Top-K optimization is already on HIVE-3562 and seemed orthogonal with this issue. And top-k will make this further less useful.
The reason why I made this is that there are so many statements like 'hive is quite inefficient handling order by cause its ran by single reducer' and I just hate hear that. This is not important issue but can be a start point for other optimization exploiting ordered traits from multiple bucket files.

Added configuration means number of reducer for lastly order-by MR stage. -1 means it will be decided by usual calculation. 0 means disabling this.
                
> Support using multiple reducer for fetching order by results
> ------------------------------------------------------------
>
>                 Key: HIVE-3972
>                 URL: https://issues.apache.org/jira/browse/HIVE-3972
>             Project: Hive
>          Issue Type: Improvement
>          Components: Query Processor
>            Reporter: Navis
>            Assignee: Navis
>            Priority: Minor
>         Attachments: HIVE-3972.D8349.1.patch
>
>
> Queries for fetching results which have lastly "order by" clause make final MR run with single reducer, which can be too much. For example, 
> {code}
> select value, sum(key) as sum from src group by value order by sum;
> {code}
> If number of reducer is reasonable, multiple result files could be merged into single sorted stream in the fetcher level.

--
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