Specifying partitioning or collecting methods
You can specify how the data is collected or partitioned before it is processed.
Partitioning data
If the stage is running in parallel mode, it processes the
data in partitions. By default, the partitioning method is set to
Auto. You can override the default behavior.
Collecting data
You can specify a collecting method.
If the stage runs sequentially, and the previous stage in
the job runs in parallel, then the data is collected before being
written. By default, the collecting method is set to Auto. You can
override the default behavior.