drill-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Serge Harnyk (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (DRILL-4436) Result data gets mixed up when various tables have a column "label"
Date Thu, 17 Mar 2016 15:56:33 GMT

    [ https://issues.apache.org/jira/browse/DRILL-4436?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15199767#comment-15199767
] 

Serge Harnyk commented on DRILL-4436:
-------------------------------------

I think the problem is more serious. I created two same tables Gender2 and Civility2 with
columns "label2" in BOTH tables.

select gender2.label2 as label2 from postgres.public.gender2 join postgres.public.civility2
on genderId = civilityId
returns: civilityLabel

select civility2.label2 as label2 from postgres.public.gender2 join postgres.public.civility2
on genderId = civilityId
returns: null

select gender2.label2, civility2.label2 from postgres.public.gender2 join postgres.public.civility2
on genderId = civilityId
returns: civilityLabel 	null

Project step works wrong when we select column with same name that in second table. Regardless
its name.

> Result data gets mixed up when various tables have a column "label"
> -------------------------------------------------------------------
>
>                 Key: DRILL-4436
>                 URL: https://issues.apache.org/jira/browse/DRILL-4436
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Storage - JDBC
>    Affects Versions: 1.5.0
>         Environment: Drill 1.5.0 with Zookeeper on CentOS 7.0 
>            Reporter: Vincent Uribe
>            Assignee: Serge Harnyk
>
> We have two tables in a MySQL database:
> CREATE TABLE `Gender` (
>   `genderId` bigint(20) NOT NULL AUTO_INCREMENT,
>   `label` varchar(15) NOT NULL,
>   PRIMARY KEY (`genderId`)
> ) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=latin1;
> CREATE TABLE `Civility` (
>   `civilityId` bigint(20) NOT NULL AUTO_INCREMENT,
>   `abbreviation` varchar(15) NOT NULL,
>   `label` varchar(60) DEFAULT NULL
>   PRIMARY KEY (`civilityId`)
> ) ENGINE=InnoDB AUTO_INCREMENT=6 DEFAULT CHARSET=latin1;
> With a query on these two tables with Gender.label as 'gender' and Civility.label as
'civility', we obtain, depending of the query :
> * gender in civility
> * civility in the gender
> * NULL in the other column (gender or civility)
> if we drop the table Gender and recreate it with like this:
> CREATE TABLE `Gender` (
>   `genderId` bigint(20) NOT NULL AUTO_INCREMENT,
>   `label2` varchar(15) NOT NULL,
>   PRIMARY KEY (`genderId`)select * from Gender
> ) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=latin1;
> Everything is fine.
> I guess something is wrong with the metadata...



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

Mime
View raw message