It is map.input.file [.start and .length also relate to the InputSplit
for the mapper]
For more: 
http://hadoop.apache.org/common/docs/r0.20.2/mapred_tutorial.html#Task+JVM+Reuse

With a custom RR, you can put in this value yourself
(FileSplit.getPath()) before control heads to the Mapper/MapRunner.

On Fri, Dec 31, 2010 at 6:17 PM, Zhou, Yunqing <[email protected]> wrote:
> You can use "map.input.split"(something like that, I can't remember..) param
> in Configuration.
> this param contains the input file path, you can use it to branch your logic
> this param can be found in TextInputFormat.java

-- 
Harsh J
www.harshj.com

Reply via email to