APD - Input Query partitioning

Hi all!

We currently designing an APD process with an input query that will deliver approx. 1M rows. On the second properties tab (Extended Setting) of the Input Query , an option is offered which should permit some kind of partitioning. Unfortunately this option is not documented. Do you have any experiences or a more detailed description of how it works ?

Any information is well appreciated.

Regards, Philipp