hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Branky Shao (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HIVE-6163) OrcOutputFormat#getRecordWriter creates OrcRecordWriter with relative path
Date Fri, 07 Mar 2014 06:10:45 GMT

     [ https://issues.apache.org/jira/browse/HIVE-6163?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Branky Shao updated HIVE-6163:
------------------------------

    Fix Version/s: 0.12.1

[~yhuai], I agree all Hive OutputFormats should follow the same way to create file path. Shall
we create another umbrella Jira for this task and fix this  this specific issue on 0.12.1?

> OrcOutputFormat#getRecordWriter creates OrcRecordWriter with relative path
> --------------------------------------------------------------------------
>
>                 Key: HIVE-6163
>                 URL: https://issues.apache.org/jira/browse/HIVE-6163
>             Project: Hive
>          Issue Type: Bug
>          Components: File Formats
>    Affects Versions: 0.12.0
>            Reporter: Branky Shao
>             Fix For: 0.12.1
>
>
> Hi,
> OrcOutputFormat#getRecordWriter creates OrcRecordWriter instance using a file with relative
path actually.
> return new OrcRecordWriter(new Path(name), OrcFile.writerOptions(conf));
> https://github.com/apache/hive/blob/7263b3bb1632b1a7c6ef5d2363e58020e1fdd756/ql/src/java/org/apache/hadoop/hive/ql/io/orc/OrcOutputFormat.java#L114
> The fix should be very simple, just as RCFileOutputFormat#getRecordWriter, append work
output path as the parent:
> Path outputPath = getWorkOutputPath(job);
> Path file = new Path(outputPath, name);
> https://github.com/apache/hive/blob/d85eea2dc5decbf23e8f4010b32f1817cf057ea0/ql/src/java/org/apache/hadoop/hive/ql/io/RCFileOutputFormat.java#L78



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

Mime
View raw message