When we talk about aggregation from an ODS to a cube what do we exactly mean? Thanks
Aneesh Gupta replied
ODS would normally have document level information but the InfoCube would have a higher level, say monthly. In that case, all documents for the given characteristics would get aggregated.
For e.g. if your ODS has Doc No., doc date, customer and material and has 10000 records for a given month. The InfoCube would have only customer, material and month. In this case the infocube will have unique rows for each customer, material and month combination. The 10000 rows for a month in the ODS should result in a lower number of rown in the InfoCube. Benefit is improved performance but dis-advantage is that an analysis lower than what has been stored in the InfoCube is not possible without going to the ODS.