Where sorting of data is done on Mapper or Reducer Node in MapReduce

[DataFlair Team]

Where sorting is done in MapReduce Job execution?

[DataFlair Team]

Sorting in MapReduce is on Both Mapper Side as well as Reducer side.

Mapper side(Quick Sort)
The Mapper process the input key value pair from Record Reader and generates output as per custom Business logic, before writing the data the output is partitioned and sorted by key.

Reducer side(Merge Sort):
In Reducer phase the data from the Mapper is again sorted as per the key and sent to Reducer for processing data as per custom business logic.The shuffling and sorting in Reducer phase happens simultaneously.

[Author]

Posts