[ 
https://issues.apache.org/jira/browse/TEZ-1317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14085305#comment-14085305
 ] 

Siddharth Seth commented on TEZ-1317:
-------------------------------------

- Following the change, all credentials setup can be removed from the examples. 
e.g. setupURIsForCredentials in IntersectExample
- On MRInput, another useful option would be .useLegacy() - so that examples 
and users who use this, can use the same configuration mechanism. The 
ShufflEdgeConfigurer also exposes this for ShuffledMergedInputLegacy
- On the exception - I was suggesting the following change.
{code}
+        throw new TezUncheckedException("inputFormat must be assignable from 
either " +
+            "org.apache.hadoop.mapred.InputFormat or " +
+            "org.apache.hadoop.mapreduce.InputFormat" +
+            " Given: " + inputFormat.getName());
{code}
s/"inputFormat must be assignable from either"/"When using this API inputFormat 
must be assignable from either .... . Otherwise use the basic configurer". 
Specifically for IOFormats which are file based, but do not extend 
FileInputFormat.
- typo : s/createDistributerDataSource/createDistributorDataSource 
(MRInputSplitDistributor)
- createGeneratorDataSource - has some unnecessary javadoc leftover from the 
previous API
- Last, changing MROutput in a similar manner would be useful (and consistent) 
- instead of exposing all the additional methods.


> Simplify MRinput/MROutput configuration
> ---------------------------------------
>
>                 Key: TEZ-1317
>                 URL: https://issues.apache.org/jira/browse/TEZ-1317
>             Project: Apache Tez
>          Issue Type: Sub-task
>            Reporter: Siddharth Seth
>            Assignee: Bikas Saha
>            Priority: Blocker
>         Attachments: TEZ-1317.1.patch, TEZ-1317.2.patch, TEZ-1317.3.patch, 
> TEZ-1317.3.patch, TEZ-1317.4.patch, TEZ-1317.5.patch
>
>
> Should at least be possible to generate the correct Descriptors.
> Potentially change the addInput / addOutput APIs to accept a single entity 
> which encapsulates InputDescriptor and InputInitializerDescriptor. Similarly 
> for Outputs.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to