During modeling I frequently need to delete some temporary tables in order to be able to complete my dataflow. This is because I am usually very close to peak consumption in memory during certain operations.
I am going to run some test operational batch dataflows on my modeling cluster and was wondering if the temporary tables are automatically deleted during the batch dataflow run, or at the end of it? Or never? I would figure it would never be never, as then memory could accumulate quite a bit over runs and prevent the batch dataflows from running.