I have a cube that contains a chart of accounts hierarchy dimension... this is of typical parent child design. The fact table contains the lowest level member in it for the aggregations to inherently sum up.
So...
DIM CHART OF ACCTS
| key | parent_key | type | account |
| 1 | NULL | SUMMARY | INCOME STMT |
| 2 | 1 | SUMMARY | SUMINCACCT |
| 3 | 1 | DETAIL | 11111111 |
| 4 | 2 | DETAIL | 22222222 |
FACT AMOUNTS
| amount | chart_key | time_key |
| 10.00 | 3 | 201211 |
| 500.00 | 4 | 201211 |
| 33.00 | 3 | 201212 |
| 5.00 | 4 | 201212 |
The problem is sometimes the users just want to slice on summary accounts and not detail. Meaning they basically don't want to see all 30,000 accounts. Only the 200 or so summary accounts when they drag and expand all on the chart of accounts hierarchy. I first thought I could just put "Type" into the dimension and they could use that as a filter. But that doesn't work because of how the fact table contains only "Types" of detail. So when slicing by summary you get nothing. Makes sense ... just missed it when originally thinking about it. My question is how can I do this... do I have to ETL it and create another key column pointing to the lowest summary account. And then create another hierarchy. Or is their a simpler more intuitive way for the user to just be able to slice on a type member like they originally wanted. Like can I create a calculated member somehow for chart of accounts based on they type attribute and the current member?
Thanks in advance...