Does the input ALWAYS need to be a FILE type ?

Hi,

I wish to pass in a large data set to hadoop's Map/Reduce "during RUNTIME" from another program for indexing purpose. Whatever code samples I have seen, invariably uses a file that they pass into JobConf...

Is it possible to just pass in a String [InputStream] into hadoop and if so how ?

Your help on this will be greatly appreciated.. Thanks

Regards
Raj