2014-09-18 7 views

답변

1
Number of mapper would be equal to number of input splits . 

when ever you submit a job ,first of all it determines number of splits ,splits are logical . 

usually one split size is equal to hdfs block size but that can also be configured where your split size could be less than or greater than block size . for efficent processing usually one split size is equal to block size . 

suppose you have file of 1 GB ,your default block size is 128 mb so approximately you will have 8 blocks and 8 input splits would be required and hence 8 mappers would be invoked for this process . 
관련 문제