[ 
https://issues.apache.org/jira/browse/CRUNCH-228?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dave Beech updated CRUNCH-228:
------------------------------

    Attachment: CRUNCH-228.patch

Here's a patch. Basically I've replaced the special case addition of the 
'.avro' extension in FileTargetImpl with a more general one. 

This did cause an integration test failure regarding Trevni files. I'm not too 
familiar with Trevni and didn't want to take a risk, so I added a change in 
TrevniKeyTarget to preserve the original behaviour rather than update the 
filenames in the test. I think changing the test would be cleaner, so let me 
know if you agree this is a better option. 
                
> FileTargetImpl cuts off extensions of output files
> --------------------------------------------------
>
>                 Key: CRUNCH-228
>                 URL: https://issues.apache.org/jira/browse/CRUNCH-228
>             Project: Crunch
>          Issue Type: Bug
>            Reporter: Dave Beech
>         Attachments: CRUNCH-228.patch
>
>
> Compressed files written by mapreduce often have extensions, e.g. '.deflate', 
> '.gz' or '.snappy'. Crunch currently cuts off these extensions during the 
> move of output files to their final destination, which is fine in some 
> circumstances but causes problems in others. 
> For example, running 'hadoop fs -text myfile.deflate' will show the 
> decompressed text on screen but running 'hadoop fs -text myfile' on a 
> deflate-compressed file with no extension prints unreadable compressed data 
> instead. 

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Reply via email to