Bikas Saha created TEZ-1389:
-------------------------------
Summary: Simplify code flow for grouping
Key: TEZ-1389
URL: https://issues.apache.org/jira/browse/TEZ-1389
Project: Apache Tez
Issue Type: Improvement
Reporter: Bikas Saha
1) If grouping is enabled then the input format is TezGroupedInputFormat
2) TezGroupedInputFormat has a config that informs it about the actual input
format (instead of the setInputFormat method that exists today).
3) Create TezInputSplit that extends InputSplit and provides more methods like
getLocationHint(). This would be useful for other use cases too.
4) When grouping is enabled then on the client the payload simply switches the
InputFormat to TezGroupedFormat and set the TezGroupedFormat wrapped format
config to the actual InputFormat. Thats it. Internally, the splits generator
will call standard config input format.
5) We can check if the InputSplit is of type TezInputSplit or not. If yes, we
can use the new methods like getLocationHint().
--
This message was sent by Atlassian JIRA
(v6.2#6252)
