Yes this is true. Combiner may never run if intermediate values don't need to shuffle out to disk before the final output is done. Also, combiner cant be substituted as a reducer.
Sent from my iPad
On Jul 10, 2011, at 4:42, Florin P <[EMAIL PROTECTED]> wrote:
> I've read on http://www.fromdev.com/2010/12/interview-questions-hadoop-mapreduce.html (cite):
> "The execution of combiner is not guaranteed, Hadoop may or may not execute a combiner. Also, if required it may execute it more then 1 times. Therefore your MapReduce jobs should not depend on the combiners execution. "
> Is it true? Also is it possible to use the Combiner without the Reducer? The framework will take into the consideration the Combiner in this case?
> Thank you,