[
https://issues.apache.org/jira/browse/HIVE-2558?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13236800#comment-13236800
]
Michael Ubell commented on HIVE-2558:
-------------------------------------
Actually I misunderstood what hive is doing. It is converting the string to a
number and then comparing it to the timestamp:
select * from rrt where r = '28801';
OK
1970-01-01 00:00:01
While this does pass the reasonability test, I don't think its the right thing
to do. It would be much better to implicitly cast the string constant to a
timestamp.
> Timestamp comparisons don't work
> --------------------------------
>
> Key: HIVE-2558
> URL: https://issues.apache.org/jira/browse/HIVE-2558
> Project: Hive
> Issue Type: Bug
> Reporter: Robert Surówka
>
> I may be missing something, but:
> After performing:
> create table rrt (r timestamp);
> insert into table rrt select '1970-01-01 00:00:01' from src limit 1;
> Following queries give undesirable results:
> select * from rrt where r in ('1970-01-01 00:00:01');
> select * from rrt where r in (0);
> select * from rrt where r = 0;
> select * from rrt where r = '1970-01-01 00:00:01';
> At least for the first two, the reason may be the lack of timestamp in
> numericTypes Map from FunctionRegistry.java (591) . Yet whether we really
> want to have a linear hierarchy of primitive types in the end, is another
> question.
--
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
