Automatic Duplicate Filter on Join Key

Question

We have a fact table and a dimension table joined normally (primary key on dimension joined to surrogate key on fact table).
One of our logical columns is using the dimension key as a source.
If we filter on the column in a report, OBIEE also filters the fact table, even though it is not a physical source of the logical column.
Is there any way to turn this behavior off? We have seen where the optimizer in 11g gets confused by the repeated value, causing it to grossly underestimate the number of rows returned. The result is a query that runs very slow.

Answer 1

Thanks, kart, but that is not the issue.
The inner join is fine, the issue is that the optimizer does not know how to handle the filter on both the fact and dimension tables.
For example,
select dim.key, sum(fact.sales)
from fact, dim
where fact.key = dim.key
and dim.key = 34
and fact.key = 34 <--- This filter is automatically added to the query, even though the fact table is not defined as a source of the logical column, key
group by dim.key;
Apparently, the optimizer in 11g gets confused by the filters on both these columns and does better when there's only one on either the fact or the dim (doesn't matter which).
I want the query to look like below but don't know how to tell OBIEE not to put the filter also on the fact.
select dim.key, sum(fact.sales)
from fact, dim
where fact.key = dim.key
and dim.key = 34
group by dim.key;