Multiple hierarchies in one dimesion

Question

Hi all,
I am new in using AWM. I have problem when using two hierarchies in one dimension. This is sample of dimension.
A
---+B
-----+B1
-------+B2
---------+B3
---+C
-----+C1
-------+C2
---------+C3
Dimension is A and it has two hierarchies B and C. Maintaining the cube is succesfull. But when "add" this cube in OBIEE and try to make request using B1 and C1 we are getting error. How we can have mulitple hierarchies in one dimension and combine columns from hierarchies? We are using database 11g, AWM 11.2.0.2 and OBIEE 10.1.3.4.
Thanks all

Answer 1

Maybe you need to join the generic dimension view A_VIEW with the fact-cube view in the physical layer of rpd on the basis of the key corresponding to the (common) lowest level of the dimension.
Regd your dimension hierarchy setup:
A
---+B
-----+B1
-------+B2
---------+B3
---+C
-----+C1
-------+C2
---------+C3It looks like you have column/level B3 as leaf level of hierarchy B and C3 as leaf level of hierarchy C.
Is this distinction piurely logical and the same physical column is used to represent both B3 and C3?
Then A_VIEW and Fact view should be joined on basis of this single common level key column.
Generally a dimension A with multiple hierarchies B and C will work fine when you have a common lowest level D, say, which acts as the leaf level for both hierarchies of the dimension.
Can you model the hierarchy as
A
---+B
-----+B1
-------+B2
---------+B3
-----------+D
---+C
-----+C1
-------+C2
---------+C3
-----------+D                             NOTE1: Same level/column D acts as leaf level of Hierarchy B as well as Hierarchy C. We should not use 2 different columns D1 and D2 even if they contain the same data
NOTE2: Dimension A is joined to Fact F1 via level/column D.
In that case, you can issue a query against Fact F1 which uses both level columns B and C from different hierarchies of A against a measure defined in Fact table F1.
You can even add a filter (where clause) on columns B/C/both B and C as required.
Result will be at the granularity of level D (not summarized to either level B along HI_B hierarchy or level C along HI_C hierarchy) -- 1 record per distinct member at Level D of default hierarchy with the corresponding values for columns B and C available as part of the output since they are ancestor fields and availble at level D. A smart reporting tool like OBIEE or BO can even exclude the column corresponding to D locally within a pivot and create ad-hoc, report level, pivot table based local summaries (aggregated results for each combination of values in columns B and C).
HTH