Thanks for the quick responds, I will definitely look into the Hadoop streaming.
What do you think about AggregationClient? It is carried out at region/region server level, maybe instead do a count/min/avg, a method can be used to write the data out to local file system?
Demai on the run
On Aug 19, 2014, at 5:04 PM, Nick Dimiduk <[EMAIL PROTECTED]> wrote: