I want to make sure that a MapReduce program (in Hadoop) does not do any combining at the mapper side. I know conf.setCombinerClass() sets the combiner class to the class which we point to.
If we don't specify the combiner class using this set function, then is it that combining is disabled or there is still some implicit default combiner applied anyways. If so, then how do we disable this combining?
There is no implicit combiner, you have to set it explicitly.
There is no default combiner, because not all map reduce algorithms even accommodate one. There's no way Hadoop could determine or generate automatically, given any arbitrary mapper and reducer classes, what combiner (if any) would even work.
The following post explains which types of MR algorithms qualify for using combiners:
http://jazzjuice.blogspot.com/2011/08/requirements-for-using-hadoop-combiner.html
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With