I've created a simple SSIS package to process my cube partitions. I set all of the Process Options to "Default" thinking this would be the most flexible way to update or fully process the cube as needed.
Using Analysis Services Processing Tasks the package first default processess the dimensions in parrallel, and then default processess a specific partition.
When I was testing the SSIS package, all tasks executed sucessfully but the entire package was finished in less than 2 seconds! I'm good but not that good, hehe. When I checked the dimension members, the daily changes from the DW had not been processed into the cube.
My question is, do I first need to process the Cube database and at what level for the daily changes to be included in the dimension and fact processing?
I've never really used the "default" option when setting up an Analysis Services processing task. As part of the SSIS job I already know that I need to do a process update on certain dimensions and/or a process full (or incremental) on a certain partition. Here is a Blog article that I read while researching this a while back that I found to be helpful. http://denglishbi.wordpress.com/2009/10/29/analysis-services-ssas-processing-and-aggregations/ as well as this one.
Hope this helps you out!
Thanks, I ended up running dynamic DDL scripts with a variable that I can control from a config file to set the process level at run time.
I also learned NOT to use the incremental process option. This option only ADDS data to your cube, think of it like appending data to a table. Not fully understanding this, I loaded duplicates into my cube in error.