Joining "semi-random" user sets in to defined OBIEE model

Question

Hi,
Where I'm currently working, I'm running in to a situation I haven't faced before. Different groups around the organization keep "lists" of employees, "lists" of customers, etc. These lists often contain not just the unique identifier for the employee / customer / whatever, but also often have associated attributes.
My challenge is this - the various groups want to join the data in those lists to already existing data in our data warehouse, using OBIEE. The big weakness here being that OBIEE has a centrally defined data model, and no really good way to allow users to join in random "stuff".
I've thought of the following ideas - none of them perfect. Has anyone faced this issue before and found a "cleaner, easier, more user friendly" way of doing this?
1. Build generic "list" tables in to the data warehouse, and allow users to upload their lists into these tables (with user ID, list name, etc). These tables would then be joined directly to the matching employee / customer / whatever tables so that OBIEE can work against them as normal.
PROS: Easy to implement and fairly easy for end users to write data to these "list" tables
CONS: Columns in the list tables, other than the "main" join column (EMPLID, for instance), all other attribute columns in the users datasources would have to be "generic". i.e. the table would have generic_char_field1, generic_num_field1, generic_date_field1, etc. If one of the attributes in the list is "employee salary" - it would have to get dumped into a field like generic_num_field1. This makes using these columns difficult to understand, and for each list the meaning of "generic_char_field1" would be something different
2. Allow users to create and upload their own tables into the DW (or their schema).
PROS: columns can be exactly what they need to be
CONS: not sure there's any way to implement this through OBIEE. Is there something in the end user layer that would allow these joins to be defined without too much work?
Any better solutions out there?
Obviously the long term solution is to get all of these "lists" and associated attributes into the DW...but that is a herculean effort for now.
Thanks,
Scott

Answer 1

Actually, what I'm hoping for is some way to write the SQL and join tables by hand vs. using the logical model. I think there are ways to bypass the business model, but I haven't used them before, and not sure if that would be within the skill of our "power users".
One other idea I had was to use BI publisher - that would make it easy to join the logical OBIEE model with custom tables (although I'm not sure BIP would understand that the tables are on the same physical database, thus the performance might suffer...).
Any other ideas? Here's the rub - we really think OBIEE is a clear winner on many other criteria...but for the ability to join in custom data sets, which some of our more vocal power users do often, it's falling really flat against vastly inferior tools (like Brio)
Thanks,
Scott