Combining relation facts with dimensions from an Essbase cube

Question

Hi!
I am having trouble combining relational measures (from EBS) with dimensions from an Essbase cube. The dimensions that we want to use for reporting (drilling etc) are in an Essbase cube and the facts are in EBS.
I have managed to import both the EBS tables and the cube into OBIEE (11.1.15) and I have created a business model on the cube. For the cube I converted the accounts dimension to a value based dimension, other than that it was basically just drag and drop.
In this business model I created a new logical table with an LTS consisting of three tables from the relational database.
The relational data has an account key that conforms to the member key of the accounts dimension in the Essbase cube. So in the accounts dimension (in the BMM layer) I mapped the relational column to correct column (that is already mapped to the cube) - this column now has two sources; the relational table and the cube. This account key is also available in the LTS of my fact table.
The content levels for the LTS in the fact table have all been set to detail level for the accounts dimension.
So far I am able to report on the data from the fact table (only relational data) and I can combine this report with account key from the account dimension (because this column is mapped to the relational source as well as the cube). But if expand the report with a column (from the accounts dimension) that is mapped only to the cube (the alias column that contains the description of the accounts key), I get an error (NQSError 14025 - see below).
Seeing as how I have modeled that the facts are connected to the dimension through the common accounts key, I cannot understand why OBIEE doesn't seem to understand which other columns - from the same dimension - to fetch.
If this had been in a relational database I could have done this very easily with SQL; something along the lines of select * from relational_fact, dim_accounts where relational_fact.account_key=dim_accounts.account_key.
Error message:
[nQSError: 14025] No fact table exists at the requested level of detail
Edit:
Regards
Mogens
Edited by: user13050224 on Jun 19, 2012 6:40 AM

Answer 1

Avneet gave you the beginnings of one way, but left out that a couple of things. First, you would want to do the export of level zero only. Second, the export needs to be in column format and third, you need to make sure the load rule you use is set to be additive otherwise the last row will overwrite the previouse values.
A couple of other wats I can think of doing this
Create a replicated partition that maps the 3 non used dimensiosn to null (Pick the member at the top of the dimension in your mapping area)
Create a report script to extract the data putting the three dimensions in the page so they don't show up.
Use the custom defined function jexport in a calc script to get what you want