[
https://issues.apache.org/jira/browse/HIVE-2508?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13132820#comment-13132820
]
Ashutosh Chauhan commented on HIVE-2508:
----------------------------------------
Thinking more about this. I dont think we want to support join on union type,
since in general real data in a union could be of any type which makes it
impossible to do comparison of union types correctly. It can be argued that in
some special cases that could be achieved via type widening, but I would argue
selecting union type for storing such data which later need to be joined is a
wrong choice. Thoughts?
If there is an agreement, I propose to disable this in semantic analysis phase
itself instead of letting it fail at runtime later.
> Join on union type fails
> ------------------------
>
> Key: HIVE-2508
> URL: https://issues.apache.org/jira/browse/HIVE-2508
> Project: Hive
> Issue Type: Bug
> Reporter: Ashutosh Chauhan
>
> {code}
> hive> CREATE TABLE DEST1(key UNIONTYPE<STRING, STRING>, value BIGINT) STORED
> AS TEXTFILE;
> OK
> Time taken: 0.076 seconds
> hive> CREATE TABLE DEST2(key UNIONTYPE<STRING, STRING>, value BIGINT) STORED
> AS TEXTFILE;
> OK
> Time taken: 0.034 seconds
> hive> SELECT * FROM DEST1 JOIN DEST2 on (DEST1.key = DEST2.key);
> {code}
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
