Hi SSAS Engineers:
This question is in the context of a complete refresh of the SSAS cube. I'm OK on the SSIS approach thanks to TMcD and past successes (also thanks to him). The question is why did I have to manually drop/delete the dimensions in BIDS SSAS, then readd them
with the cube wizard to pull in the dimensions with the same names, but new surrogate key relationships and more surrogate keys in total?
We need to Automate in SSIS Cube Drop and Recreate of Dimensions. I originally had to perform this function on the Solution Explorer Manually in SSAS UDM, since the names are the same, but the structures and the data are changing for the dimensions
on the refresh (and the fact table is subsequently increasing). Herein lies the problem.
We had a keynotfound error in the cube processing, since the old keys were being kept. I had run Full processing on the dimensions, but since the dimensions have the same name (no drop old names and add new names), the old surrogate keys remained. The fact
table had the old row count as well, since the surrogate key pipeline to look up the surrogate keys for the new dimensions only had the old dimensions, even though I truncated the fact table. Yes--I had already rebuilt the dimensions in SQL Server and the
row counts were the new refresh row counts.
The dimension counts were right according to refresh numbers after
View Complete Post