Where Connect for Big Data Sort Fits into MapReduce - Connect_ETL - 9.13

When a MapReduce job is invoked with Connect for Big Data sort acceleration, Connect for Big Data Sort replaces the following stages of the job processing:

• Sort (on the map side), if one or both of the following conditions is true:
• the property dmx.map.sort.a is set to ‘d’
• the map output compression is set to bzip2
• Otherwise, the “null sort” is invoked, in which case Connect for Big Data partitions but does not sort the data on the map side.
• Merge (on the reduce side)