How to globally sort the output from a Mapreduce Job?

TotalOrderPartitioner class can be used instead of default HashPartitioner class to generate the output from Mapreduce job to be in globally sorted order.

Job.setPartitionerClass(TotalOrderPartitioner.class);