Reputation: 47
i want to know how many files will be produced if only a single mapper ( no reducer,no combiner etc ) is run for all file splits .
example- if there are 4 file splits . then there is single mapper that will process all file splits. how many files as mapper output ?? -> one or four
Upvotes: 1
Views: 475
Reputation: 1411
Each map task will produce one output file. If you have one file on HDFS that is split into four blocks, you will get four output files from a Map-Only job. If the input file is not in a splittable format, like GZip, it will be combined and only one mapper will act on it, outputting one file.
Upvotes: 4