hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tom White (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-1936) WordCount unit test plus a helper class to facilitate testing Mappers and Reducers
Date Tue, 09 Oct 2007 20:43:50 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-1936?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12533496
] 

Tom White commented on HADOOP-1936:
-----------------------------------

Sorry to come to this so late, but I would vote to use a mock object library (jMock) to write
tests for mappers and reducers as you argued for on the user list (http://www.mail-archive.com/hadoop-user@lucene.apache.org/msg02261.html).
Using jMock is much more concise than the approach in this patch (see http://www.matthewekent.com/?p=15),
although I admit it does have a bit of a learning curve.

> WordCount unit test plus a helper class to facilitate testing Mappers and Reducers
> ----------------------------------------------------------------------------------
>
>                 Key: HADOOP-1936
>                 URL: https://issues.apache.org/jira/browse/HADOOP-1936
>             Project: Hadoop
>          Issue Type: Test
>          Components: examples
>            Reporter: Kate Rhodes
>            Priority: Trivial
>         Attachments: WordCountTestExampleAndOutputCollectorTestHelper.patch
>
>
> There are to pieces to this The first is a test for WordCount, not because word count
actually needed one but because it would be useful to beginners to have an example of how
to unit test Mappers and Reducers.
> The second piece is AOutputCollector and it's associated unit test TestAOutputCollector.
This is an abstract class that can be quickly extended by a stub OutputCollector in your unit
tests to collect the output from your Mapper and Reducer tests and make it available for easy
retreival when testing to see if the fourth key that was emitted was the one you expected.
 I think that this would be a useful tool to have in the main test folder but wasn't sure
where would be best to put it. Also, since nothing else in Hadoop uses Hungarian notation
you'll probably want to rename it. I didn't because I'm not confident about the naming conventions
here and figured that since it  and its test probably wouldn't end up living in the same folder
as WordCount that you could just rename it when you moved it.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message