I'm setting up the following to handle our very large daily data loads into a dimensional database and would appreciate any advice or comments if I'm doing something that could cause me problems, or an improvement could be made.
In the RdB on the same server as we have our dimensional dB, I have factTable_currentDate with only the current days' data (approx. 14M records). The remaining data is kept on a separate server, partitioned by month. However I'm hoping I will rarely, if ever, need to access them when updating the cube.
In the SSIS package I will:
1. Truncate then import the current days' data in the rDb factTable_currentDate
2. Process Update the changing dimensions
3. Process Incremental the current days' fact table
Because each days' data is standalone i.e. there are no updates/deletes/additions to prior days' data, I'm thinking this will not cause me any issues. Same with the dim tables - any new attributes will only affect fact data going forward.
WIll this work as I'm hoping and let me do miminal processing to keep the cube updated?