Hi. I read the post at http://social.msdn.microsoft.com/Forums/en-US/sqlanalysisservices/thread/d46a292a-9f1e-4682-a7aa-e3cd99b97f6b but am not convinced.
We run std 2008 (there is no partitioning in std 2008).
Typically we do one of two things on our larger cube for negating facts. We either insert offsetting records into our star prior to the next incremental or we delete the records from the star and process full.
But in a much smaller cube that we have, I'd like to investigate other options due to the trial and error nature of finalizing a new month's accounting entries in this smaller cube. I dont want to process full every time. I dont like the noise of "offsetting" records in an app like this.
Isnt it true that there is a direct insert/update/delete feature available on molap if a certain setting is made correctly? If yes, please expound. We may be open to going enterprise for reasons other than its partitioning capabilities.
Also/alternatively, if i was willing to go holap or rolap on this app, might deletion of facts be easily accomplished by my simply deleting facts from the source sql table?