After image delta from cube to OHD..

Question

Hi All,
I would like to pick you brains for an issue I am currently facing. There is one requirement that requires us to send delta data in the form of After image to an OHD file from a cube.
The issue is when we get delta data from the cube, we are getting only the additive values. My understanding is that this delta is the aggregation of the Before and After image delta values from the underlying dso.
My issues are:
1. Is there a way of getting the after image delta values from the cube?
2.  I do not want to read from the underlying dsos because, some calculations are performed when loading the cubes. Also, the data required is in aggregated on year level. So if we take the data from the dso, which is in day level, we would have to do aggregation manually.
Thanks in advance for your help experts!

Answer 1

Thanks for the above responses...
Let me give the more detailed scenario..
I have 4 GL cubes whose delta values are required to be exported. They want the delta values to reflect the "As Is",
So for instance we have the following on Day 1:
GL Acc  0Amount
10001    100
The next day, 0Amount(Aggregation mode: Sum) gets an addition of 100, they want to see the following
GL Acc 0Amount
10001   200
These 4 GL Cubes have about 5 common underlying DSOs.
The issues here is:
- If we read the delta directy from Cubes we are getting the addiditive values: in this case +100
- If we read the delta from DSOs, from my understanding we should perform aggregation manually. Since the client wants to view the data by Year instead of by Day which is how it exists in the DSO.
- Another issue here is: The clients require us to send a SINGLE file. Hence we need a way of appending the Data from multiple InfoProviders to a single file.
Our Proposed Solution:
- Read from underlying DSOs. and store in a intermediate Write - Optimized DSO.
- Perfrom manual aggregation by using ABAP.
- Send delta values to OHD filtering out the before image(ORecordMode = 'X')
Issues:
Need to perform manual aggregation which may be an issue because the volume can be large.(Currently we can not predict the volume.)
Plus Any other issues that you can foresee?