Mailing-List: contact dev-help@hive.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hive.apache.org
Date: Mon, 19 Sep 2011 15:00:10 +0000 (UTC)
From: "Chinna Rao Lalam (JIRA)" <jira@apache.org>
To: hive-dev@hadoop.apache.org
Message-ID: 
 <370750865.42366.1316444410717.JavaMail.tomcat@hel.zones.apache.org>
In-Reply-To: 
 <625079731.25780.1313017948931.JavaMail.tomcat@hel.zones.apache.org>
Subject: [jira] [Commented] (HIVE-2367) Indexes' storage descriptors'
 columns are not deduped, and altering an index leaves behind an unused
 storage descriptor
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HIVE-2367?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13107879#comment-13107879 ] 

Chinna Rao Lalam commented on HIVE-2367:
----------------------------------------

@Sohan:
I am trying to fix this issue but i am not able to reproduce this issue. I ran the below queries

{quote}
create table grade (userid string,name string,maths int,physics int,chemistry int) ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t';

load data local inpath '/home/chinna/data/grade1.txt' into table grade;

CREATE INDEX grade_index ON TABLE grade(name)as 'compact' WITH DEFERRED REBUILD;
  
ALTER INDEX grade_index ON grade REBUILD;
{quote}

After the executioon of above queries 3 storage descriptors are created.
After the execution of above queries, if ALTER INDEX query is executed the storage descriptors number is not increased.

Pls check this and post ur queries.

> Indexes' storage descriptors' columns are not deduped, and altering an index leaves behind an unused storage descriptor
> -----------------------------------------------------------------------------------------------------------------------
>
>                 Key: HIVE-2367
>                 URL: https://issues.apache.org/jira/browse/HIVE-2367
>             Project: Hive
>          Issue Type: Bug
>          Components: Metastore
>            Reporter: Sohan Jain
>
> The metastore migration by HIVE-2246 does not dedupe the COLUMNS information for Indexes.  That is, the IDXS table has a Storage Descriptor that always points to a new Column Descriptor, which is unlikely shared by any other storage descriptor.
> Therefore, when altering an index, a new storage Descriptor and column descriptor are created.  No other objects will reference the old storage descriptor and column descriptor, but they will persist in the metastore db.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira