hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gopal V (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-8078) ORC Delta encoding corrupts data when delta overflows long
Date Mon, 22 Sep 2014 23:00:38 GMT

    [ https://issues.apache.org/jira/browse/HIVE-8078?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14143991#comment-14143991
] 

Gopal V commented on HIVE-8078:
-------------------------------

LGTM - +1 on changes, pending tests.

> ORC Delta encoding corrupts data when delta overflows long
> ----------------------------------------------------------
>
>                 Key: HIVE-8078
>                 URL: https://issues.apache.org/jira/browse/HIVE-8078
>             Project: Hive
>          Issue Type: Bug
>          Components: File Formats
>    Affects Versions: 0.14.0, 0.13.1
>            Reporter: Tim Patterson
>            Assignee: Prasanth J
>            Priority: Critical
>         Attachments: HIVE-8078-testcase.patch, HIVE-8078.1.patch, HIVE-8078.2.patch,
HIVE-8078.3.patch, HIVE-8078.4.patch
>
>
> There is an issue with the integer encoding that can cause corruption in certain cases.
> The following 3 longs cause this failure.
> 4513343538618202711
> 2911390882471569739
> -9181829309989854913
> I believe that even though the numbers are in decreasing order, the delta between the
last two numbers overflows causing a positive delta, in this case the last digit ends up being
corrupted (the delta is applied for the wrong sign resulting in -3442132998776557225 instead
of -9181829309989854913.



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

Mime
View raw message