hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ashutosh Chauhan (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-7095) Fix test fails for both hadoop-1 and hadoop-2
Date Thu, 22 May 2014 06:54:39 GMT

    [ https://issues.apache.org/jira/browse/HIVE-7095?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14005675#comment-14005675
] 

Ashutosh Chauhan commented on HIVE-7095:
----------------------------------------

I am not sure for problem 2, if there is any advantage of writing different versions of test
for hadoop-1. Clearly, test is written with assumption that it will generate 2 files, which
is not possible with # of reducers = 1 which is what we get with hadoop-1. So, test is not
well suited for hadoop-1, I think we should just exclude it for version < hadoop-2, instead
of having 2 test variant where obviously one is wrong. Ofcourse, if you want we can keep 2
versions, but IMO its not serving any purpose. 

> Fix test fails for both hadoop-1 and hadoop-2
> ---------------------------------------------
>
>                 Key: HIVE-7095
>                 URL: https://issues.apache.org/jira/browse/HIVE-7095
>             Project: Hive
>          Issue Type: Task
>          Components: Tests
>            Reporter: Navis
>            Assignee: Navis
>         Attachments: HIVE-7095.1.patch.txt, HIVE-7095.2.patch.txt
>
>
> There are some try to fix test results for hadoop-2(HIVE-6838, HIVE-6979, etc.). But
these remained problems for hadoop-1. Let's finish this here, not prolonging output ping-pongs
between two versions.
> 1. Data size in explain result is different
> bucketizedhiveinputformat.q
> 2. forced bucketing(=2) does not make two files in hadoop-1. 
> groupby_sort_1.q
> groupby_sort_skew_1.q
> 3. Column stats of srcpart, which has loaded a file rather than inserted, has COLUMN_STATS_ACCURATE=true,
numRows=0, rawDataSize=0 for hadoop-2, which seemed wrong.
> sample8.q
> transform_ppr1.q
> transform_ppr2.q
> union_ppr.q
> 4. Iteration order of table properties for show create table is different.
> nullformatCTAS.q
> show_create_table_alter.q
> 5. Iteration order of show table properties is different
> unset_table_view_property.q



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message