Shuffling and selection are two key processes that operate simultaneously during mapper and reducer operations.
Shuffling is the procedure to move data from Mapper to reducer. It is an obligatory action for reducers to continue work, as the shuffling process is a contribution to the reduction duties.
In MapReduce, the key value pairs between the map are automatically sorted and phases reduced (after the mapper) before travelling to the reducer. In applications where you need to sort at some phases, this functionality is beneficial. It also spares the whole time of the programmer.