Streaming should allow TextInputFormat keys to be passed through
----------------------------------------------------------------
Key: MAPREDUCE-1350
URL: https://issues.apache.org/jira/browse/MAPREDUCE-1350
Project: Hadoop Map/Reduce
Issue Type: Improvement
Components: contrib/streaming
Affects Versions: 0.22.0
Reporter: Todd Lipcon
Streaming's PipeMapper automatically ignores the key (LongWritable file offset)
coming from TextInputFormat. This is usually what the user wants, but
occasionally the file offsets are useful - for example, after grepping a large
file for a particular pattern, you may want to look at the offset in the file
where the pattern matched. There should be a boolean configuration variable
which overrides ignoreKey in PipeMapper.java:86
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
