You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@trafodion.apache.org by "Suresh Subbiah (JIRA)" <ji...@apache.org> on 2015/10/08 00:48:26 UTC

[jira] [Updated] (TRAFODION-1164) LP Bug: 1443463 - '\N' and empty string are both treated as null when selecting hive table.

     [ https://issues.apache.org/jira/browse/TRAFODION-1164?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Suresh Subbiah updated TRAFODION-1164:
--------------------------------------
       Assignee: Selvaganesan Govindarajan  (was: Apache Trafodion)
    Component/s: sql-exe

> LP Bug: 1443463 - '\N' and empty string are both treated as null when selecting hive table.
> -------------------------------------------------------------------------------------------
>
>                 Key: TRAFODION-1164
>                 URL: https://issues.apache.org/jira/browse/TRAFODION-1164
>             Project: Apache Trafodion
>          Issue Type: Bug
>          Components: sql-exe
>            Reporter: Howard Qin
>            Assignee: Selvaganesan Govindarajan
>            Priority: Critical
>              Labels: hive, null
>             Fix For: 2.0-incubating
>
>
> When loading/selecting from hive table,  we treat both empty string and '\N' as null, this doesn't conform with hive behavior, in which, user can configure what null character is.
> The issue can be explain in following steps:
> 1)
> >select * from t1;
> A            B
> -----------  ------------------------------------------------------------------------------------------------------------------------
>           4
>           1  ?
>           2  fsdf
>           3  hhah
> 2)
> unload  with null_string '\N' into '/nulltest' select * from t1; 
> -bash-4.1$ swhdfs dfs -cat /nulltest/file0-20150410015053-828
> 4|
> 1|\N
> 2|fsdf
> 3|hhah
> 3)
> hive> create external table nulltest ( A int, B string ) 
>          >                                               row format delimited fields terminated by '|'
>          >                                                location '/nulltest'
>          >                                                tblproperties ('serialization.null.format' = '\\N');
> hive> select * from nulltest;
> OK
> nulltest.a      nulltest.b
> 4                                                       <--  empty string
> 1                       NULL
> 2                       fsdf
> 3                       hhah
> 4)
> >>select * from hive.hive.nulltest;
> A            B
> -----------  ------------------------------------------------------------------------------------------------------------------------
>           4  ?                               <--  empty string is considered null. 
>           1  ?                                  
>           2  fsdf
>           3  hhah
> >select * from t1;
> A            B
> -----------  ------------------------------------------------------------------------------------------------------------------------
>           4                            <-- should consist with t1 and hive.
>           1  ?
>           2  fsdf
>           3  hhah
> Assigned to LaunchPad User khaled Bouaziz



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