Hi,
I hope this is sent to the right forum.
This is a question where I believe that the answer is "no - not possible" but I would really like it to be "yes it is possible".
Can anyone confirm my assumptions? Does anyone else feel this should be on a wish-list?
First some words on the setup.
I have two dataproviders. DP1 and DP2. The source data are from different universes as the datasources are on different servers.
The dataproviders have multiple non-equal dimensions, so linking the datsets is not an option. To effectively linking two datasets you must link all dimensions from one of the datasets to the other.
In some (many?) cases I think this is overly strict.
Say we build a clever universe/query so that the DP1 query has a dimension which is essentially the PK of the dataset retreived. This would require only that dimension to be linked and the rest of the data would be fine.
One option I've been considering is to make all dimensions in the universe as "details" under that "PK" of the retreived dataset. However that would mess up the dimensional hierarchy that is setup in the universe (i.e. an object is either a detail or a dimension / or you need duplicate definitions of ALL dimensions and you really need to keep track of what you're doing).
To generalize - in other setup we could have the data organized in granularity hierarchies. Assuming straight hierarchies we would only need to link the most granular detail of each such hierarchy in order to link two data providers.
/regards Mats