Hi All
In BW on HANA I have created a Non-Cumulative cube with 1 simple Non-cumulative key figure and several cumulative key figures. I then have 2 simple queries which I am checking via RSRT (so removing additional network latency that could be added if utilizing BEx or Analysis for Office)
In Query #1 - I have only the cumulative KFs and the response time for returning large volumes of data are excellent.
In Query #2 - I remove all cumulative KFs and just have the 1 Non-*** KF. This time the response time is very noticeably longer.
Checking the query SQL for #2 and executing this directly on HANA, I can see this is executed in microseconds so the large time difference is obviously when the information is being passed back to the "front-end" - as I mentioned in this case RSRT.
My theory is that due to the nature of a Non-*** KF with the fact that it has an inflow and outflow key figure, the HANA engine is not currently performing this aggregation, but I am not sure why or how to force it do so? Have ensured that there is enough memory in terms of head-room available so it is not blowing space limits.
Has anyone else come across similar examples where the Non-*** KF has a significant affect on the response times?
Any ideas how to improve this situation?
Kind regards