Perfomance vs statistics

Ever wondered why, even though you've deleted all indexes on your cube, the (massive) initial load still took a very, very long time?
It's all about (database) statistics!
Unfortunately the statistics can't be refreshed during (massive) data loading...even though you might think it's possible, looking at the screenshot below.

Unfortunately, the "refresh statistics" button doesn't work properly.

Luckily there's a workaround: Refreshing the statistics (of a cube), while data is being loaded towards it.
Run report RSANAORA and enter the FACTtable of the cube for which the statistics needed to be refreshed:



ps. With transaction DB20 you can validate/check the correctness of the current statistics. As long as the traffic light is green, you've got nothing to worry about :-)

Starting a DTP WITHOUT a process chain

In the old days, working with infopackages, it was easy to schedule loads from within the infopackage. Schedule the infopackage directly, or in the background on a specific date + time.
Unfortunately this easy scheduling is not available from within the DTP. To schedule an individual DTP, you HAVE to create a process chain, containing the DTP you want to schedule.

Luckily there's a workaround for this. Applying the code below will create an ABAP which makes it possible to schedule (individual) DTP's, without the necessity of a process chain.



Enter the technical DTP name in the parameter field of the ABAP and schedule the ABAP. Via scheduling the ABAP automatically the DTP will be scheduled ;-)