[
https://issues.apache.org/jira/browse/MRUNIT-66?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13479319#comment-13479319
]
Carlos Espinoza commented on MRUNIT-66:
---------------------------------------
I'm having a problem with this change. Perhaps you can suggest a work around.
We are currently upgrading to CDH4 which required us to use MRUnit 0.9.0
It doesn't seem like this fix reflects what the framework does. Hadoop allows
for a null key when it's not being used as an intermediate key. For instance,
if the key is null and the value is something else. Then it'll only write the
value to the output and it doesn't complain. It only has a problem with it when
it is being used as an intermediate key
Having said that, some of our mapper tests are failing with returnNonNull when
they output null as a key even though they are not being used as intermediate
keys. The last thing we want to do is change the mapper code, so I'm wondering,
are there any work arounds you can suggest? Thanks for your help
> null input checks and behavior on no input to a driver are inconsistent
> -----------------------------------------------------------------------
>
> Key: MRUNIT-66
> URL: https://issues.apache.org/jira/browse/MRUNIT-66
> Project: MRUnit
> Issue Type: Improvement
> Affects Versions: 0.8.1
> Reporter: Jim Donofrio
> Assignee: Jim Donofrio
> Priority: Minor
> Fix For: 0.9.0
>
>
> the MapDriver class does not allow null input values if using setInput(Pair)
> but does allow null input if using setInputKey, setInputValue, or
> setInput(key, value)
> Also the MapDriver, ReduceDriver classes will throw null pointer exceptions
> with no input while the MapReduceDriver and Pipeline classes will just log
> warnings
--
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
