[
https://issues.apache.org/jira/browse/HIVE-4928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13722076#comment-13722076
]
Jason Dere commented on HIVE-4928:
----------------------------------
Review at https://reviews.facebook.net/D11871
These changes make use of a Java action in ANTLR grammar to create a
DATELITERAL token containing the text of the date literal string. There didn't
seem to be any other way to do that.
Ran unit tests yesterday night on Mac, 4 failed, but these all passed when I
ran them on a Linux VM.
> Date literals do not work properly in partition spec clause
> -----------------------------------------------------------
>
> Key: HIVE-4928
> URL: https://issues.apache.org/jira/browse/HIVE-4928
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Reporter: Jason Dere
> Assignee: Jason Dere
> Attachments: HIVE-4928.1.patch.txt, HIVE-4928.D11871.1.patch
>
>
> The partition spec parsing doesn't do any actual real evaluation of the
> values in the partition spec, instead just taking the text value of the
> ASTNode representing the partition value. This works fine for string/numeric
> literals (expression tree below):
> (TOK_PARTVAL region 99)
> But not for Date literals which are of form DATE 'yyyy-mm-dd' (expression
> tree below:
> (TOK_DATELITERAL '1999-12-31')
> In this case the parser/analyzer uses "TOK_DATELITERAL" as the partition
> column value, when it should really get value of the child of the DATELITERAL
> token.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
