hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sergey Shelukhin (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-8619) CBO causes some more type problems
Date Tue, 28 Oct 2014 22:12:41 GMT

    [ https://issues.apache.org/jira/browse/HIVE-8619?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14187577#comment-14187577
] 

Sergey Shelukhin commented on HIVE-8619:
----------------------------------------

I understand why this would work (in fact I was going approximately the same way, making the
parameter true for some extra cases for double :)) but what is the deeper reason for treating
double differently?
Perhaps literal should be number-formatted differently to string instead in case when it's
not explicit?

> CBO causes some more type problems
> ----------------------------------
>
>                 Key: HIVE-8619
>                 URL: https://issues.apache.org/jira/browse/HIVE-8619
>             Project: Hive
>          Issue Type: Bug
>          Components: CBO
>            Reporter: Sergey Shelukhin
>            Assignee: Laljo John Pullokkaran
>            Priority: Critical
>         Attachments: HIVE-8619.patch
>
>
> On recent rerun, tests such as char_cast, decimal_2, literal_double, vector_elt (incl.
in Tez), udf4, vector_decimal_math_funcs have changes in results it looks like something is
broken with types again; ".0" suffix disappeared in some cases (in char_cast it's probably
the simplest to see), number formats changed when casts are involved.
> Also, windowing_expressions and couple more tests fail due to attempts to cast Double
to Int writables, might be the same issue.



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

Mime
View raw message