We currently have an input query based on an aggregation level that is based on a multiprovider (which does not include any aggregation levels) that combines four cubes one of which is a real-time infocube. We are running the query with BEx in Excel, and it is extremely slow. As an experiment, I completed a query copy and based it on the multiprovider instead of the aggregation level. It appears to work fine allowing editing and saving back to the realtime cube and runs in under a minute. However, everything I have read says that an input query must be based on an aggregation level or a multiprovider that includes at least one simple aggregation level. What are the consequences of using the query based on the multiprovider? Has anyone else encountered this?