I have a data cube in Analysis Services (SQL Server 2000) with private dimensions only. The cube uses HOLAP storage and the dimensions use MOLAP storage. I have been testing loading and processing the cube and have noticed that when I clear all records out of the fact table and then reload them, process the cube using a full process after the first 100,000 test records and an incremental processing after the second 100,000 records (for a total of 200,000 records in the fact table), the total record count is 400,000. Obviously, when I cleared out the fact table, brought in the test records, the original 200,000 test records that were in the cube prior to bringing in the second test batch did not clear out of the aggregations.
When using the full processing method, doesn't that clear out the aggregations and start over again? If not, what is needed to completely clear out the data cube?
Thank you,