Hi SAP Experts,
Recently we have converted our huge InfoCube to an SPO (Semantically Partitioned Object). This is part of our exercise to reduce the footprint of BWA Index for our InfoCube thus will reduce the time taken to process BWA Index. Apparently after this our BW Query has been performing very slow.
While studying our execution log by turning on statistics in RSDDSTAT, we noticed that it has been reading all individual partitions when the query is executed. E.g. we have a total of 10 partitions with each partitioned based on 0CALDAY. Sad to say 1 of our BW Query has the selection on another type of Fiscal Period (not the regular 0FISCPER) E.g. but instead another set of Fiscal Period called E_FISCPER.
What we have observed is time taken to return data for this query seems to multiply by the number of partitions it exists. In this case 10 x 20 secs = 200 seconds. I am wondering in BWA Index, is there an option to set some parallelization to read simultaneously all partitions? Or could it be some settings in SPO that we need to explore for this parallelization to happen?