[ https://issues.apache.org/jira/browse/CRUNCH-146?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dave Beech updated CRUNCH-146:
------------------------------
Attachment: CRUNCH-146.patch
Here's a patch to remove the file extension checking.
If it turns out checking for '.avro' filenames is the correct thing to do, can we instead
have loud warnings or a job failure if no valid input paths are provided?
> Jobs given Avro input data without '.avro' extension succeed, but do nothing
> ----------------------------------------------------------------------------
>
> Key: CRUNCH-146
> URL: https://issues.apache.org/jira/browse/CRUNCH-146
> Project: Crunch
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.4.0
> Reporter: Dave Beech
> Assignee: Josh Wills
> Attachments: CRUNCH-146.patch
>
>
> The Crunch AvroInputFormat class explicitly checks each input path for a '.avro' extension.
> If none of the input paths have such an extension, the job continues and succeeds even
though no map tasks get allocated, and none of the pipeline stages are executed.
> ( I notice this behaviour is taken from Avro's own input format. It does happen in standard
Avro mapreduce too therefore, but only in the old mapred API - the new mapreduce API version
will happily read files regardless of extension. I'm going to report this as a bug on Avro
as well...)
--
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
|