It has been some time since we first started using personalization extensions for analysis services.
See: Personalization Extensions
They are quite powerful, and the users who have discovered them are now building more and more custom members, for their ad-hoc reporting purposes. The primary types of items that users want are personalized sets, and calculated members. These are introduced into the cubes during the "CubeOpened" event by calling "CREATE SET" and "CREATE MEMBER".
The problem we are experiencing nowadays is that users are taking this feature further than what I had expected. They are using them to create hundreds of ad-hoc custom sets and members. When these users connect to the cubes in their Excel workbooks, it can take 30-60 seconds for the SET's and MEMBER's to get created. This means they have to wait a long time before they can even start making queries against their pivot tables.
Moreover, there doesn't seem to be a way to create more than one at a time. They have to be requested individually , via AdomdCommand ( Link ). I've found that as you keep calling CREATE MEMBER, the operations get increasingly slower (ranging from 1 ms each at the beginning to about 30 ms each after 500 members have already been created).
Can anyone tell me if they've used personalization extension to this extent, or encountered this scalability problem when using lots of SET's and MEMBER's? I'd really like to be able to create a large number of custom calculations (via CREATE MEMBER) without getting bogged down. As near as I can tell, the programming interface doesn't allow the bulk creation of multiple members, nor does it allow a personalized calculation script. Any pointers would be greatly appreciated.
David Beavon