DSO Keyfields for 0CO_OM_CCA_1 datasource

Question

Hello All,
I am using datasource 0CO_OM_CCA_1 to load my plan data. My most important concern is clearing out of the full load when I load the DTP from this DS as its full. I saw some suggestions here to use a Z DSO to avoid complications and just upload deltas. In that case what should be my key fields for the custom DSO.
Please advise.
Much thanks

Answer 1

Hi Krrish,
I would like to bring to your notice the benefits and limitations of the two stratergies, which may be would help you chose the correct one for your scenario.
Deletion of overlapping request -
1. If you choose delete when the selections are same or more comprehensive, any repair load would need to have the same granularity in selection as your original full DTP, else you run the risk of loosing data. ie : If you are loading period wise, weekly full loads to your cube for that corresponding period or week(selection of period and week).
For any repair load you would do, you would have to use the selection of the granularity of period or week or a completely disparate selection.
If you use a selection of period and CO area say, then you will delete a request in the cube for the same period and having data of all CO areas.
An example of diosparate selection would be, use CO Area alone as the selection and use selective deletion of data all periods and then load this CO area for all periods. But again, here you have a risk that next time you do a repair for any past preiod, this DTP with disparate selection does not get deleted and you will end up with duplicates and again a selective deletion would be required.
Summary - Repairing is complicated and needs data sets to be loaded in terms of full week or period
2.You cannot compress the requests... otherwise delete overlapping selections would not work...Unless you surely are not going to load the data of the request being compressed.
Advantages - Can be automated using process chain as you are already aware...
DSO before Cube
1. You reduce the data flow to the cube, by allowing only deltas to flow through. The amount of data being loaded and deleted is lesser, thus improving load performance in the cube. But you would increase activation time in the DSO.
2. Repair from R/3 is simple in this case. Load the repair load to DSO and a delta generated would correct the data.
3. Cube can be compressed. Improving query performance.
Hope it helps,
Best regards,
Sunmit.