Mailing-List: contact issues-help@hive.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hive.apache.org
Date: Fri, 31 Jul 2015 07:16:04 +0000 (UTC)
From: "Matt McCline (JIRA)" <jira@apache.org>
To: issues@hive.apache.org
Message-ID: <JIRA.12850217.1438255611000.337847.1438326964533@Atlassian.JIRA>
In-Reply-To: <JIRA.12850217.1438255611000@Atlassian.JIRA>
References: <JIRA.12850217.1438255611000@Atlassian.JIRA>
 <JIRA.12850217.1438255611552@arcas>
Subject: [jira] [Commented] (HIVE-11410) Join with subquery containing a
 group by incorrectly returns no results
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HIVE-11410?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14648855#comment-14648855 ] 

Matt McCline commented on HIVE-11410:
-------------------------------------

[~nbrenwald] please attach your EXPLAIN plan for the query.  And, confirm you are using branch-1.1

> Join with subquery containing a group by incorrectly returns no results
> -----------------------------------------------------------------------
>
>                 Key: HIVE-11410
>                 URL: https://issues.apache.org/jira/browse/HIVE-11410
>             Project: Hive
>          Issue Type: Bug
>          Components: Hive
>    Affects Versions: 1.1.0
>            Reporter: Nicholas Brenwald
>            Assignee: Matt McCline
>            Priority: Minor
>         Attachments: hive-site.xml
>
>
> Start by creating a table *t* with columns *c1* and *c2* and populate with 1 row of data. For example create table *t* from an existing table which contains at least 1 row of data by running:
> {code}
> create table t as select 'abc' as c1, 0 as c2 from Y limit 1; 
> {code}
> Table *t* looks like the following:
> ||c1||c2||
> |abc|0|
> Running the following query then returns zero results.
> {code}
> SELECT 
>   t1.c1
> FROM 
>   t t1
> JOIN
> (SELECT 
>    t2.c1,
>    MAX(t2.c2) AS c2
>  FROM 
>    t t2 
>  GROUP BY 
>    t2.c1
> ) t3
> ON t1.c2=t3.c2
> {code}
> However, we expected to see the following:
> ||c1||
> |abc|
> The problem seems to relate to the fact that in the subquery, we group by column *c1*, but this is not subsequently used in the join condition.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)