hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sandy Ryza (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-5457) Add a KeyOnlyTextOutputReader to enable streaming to write out text files without separators
Date Mon, 14 Oct 2013 20:41:43 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-5457?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13794455#comment-13794455
] 

Sandy Ryza commented on MAPREDUCE-5457:
---------------------------------------

Uploading a patch with a more comprehensive testcase.  Discovered in the process of writing
it that streaming didn't allow you to specify an output reader independently of an input writer.
 So I added that capability as well.

> Add a KeyOnlyTextOutputReader to enable streaming to write out text files without separators
> --------------------------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-5457
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5457
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>    Affects Versions: 2.1.0-beta
>            Reporter: Sandy Ryza
>            Assignee: Sandy Ryza
>         Attachments: MAPREDUCE-5457-1.patch, MAPREDUCE-5457-branch-1.patch, MAPREDUCE-5457.patch
>
>
> MR jobs sometimes want to just output lines of text, not key/value pairs.  TextOutputFormat
handles this by, if a null value is given, outputting only the key with no separator.  Streaming
jobs are unable to take advantage of this, because they can't output null values.  A text
output format reader takes each line as a key and outputs NullWritables for values would allow
streaming jobs to output lines of text. 



--
This message was sent by Atlassian JIRA
(v6.1#6144)

Mime
View raw message