[jira] [Updated] (HIVE-8484) HCatalog throws an exception if Pig job is of type 'fetch'

["Lorand Bendig (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HIVE-8484?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Lorand Bendig updated HIVE-8484:
--------------------------------
    Attachment: HIVE-8484.patch

As a solution I'd check for the property 'pig.job.converted.fetch' and if set, would return true.

> HCatalog throws an exception if Pig job is of type 'fetch'
> ----------------------------------------------------------
>
>                 Key: HIVE-8484
>                 URL: https://issues.apache.org/jira/browse/HIVE-8484
>             Project: Hive
>          Issue Type: Bug
>          Components: HCatalog
>    Affects Versions: 0.14.0
>            Reporter: Lorand Bendig
>             Fix For: 0.14.0
>
>         Attachments: HIVE-8484.patch
>
>
> When Pig tries to retrieve result in fetch mode through HCatalog then HCatLoader#setLocation(String location, Job job) can't set the outputschema because HCatUtil#checkJobContextIfRunningFromBackend(job) always returns false :
> {code}
> public static boolean checkJobContextIfRunningFromBackend(JobContext j) {
>     if (j.getConfiguration().get("mapred.task.id", "").equals("") &&
>         !("true".equals(j.getConfiguration().get("pig.illustrating")))) {
>       return false;
>     }
>     return true;
>   }
> {code}
> This is because in fetch mode we don't have a mapred.task.id. A null outputschema will raise an exception when HCatBaseLoader#getNext() is called: (ERROR 6018: Error converting read value to tuple).



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