Aggr dimensions and global selection

[#1]

Hi,

I have many fact tables, with composite keys intersecting each other in different ways.

I used Link Table design, so I have a common "Keys" table linking each fact.

Each fact table has a data field, so a data selection on one of them results in a selection on "Keys" linktable, and so on each other fact.

My pivot's calculated dimension is

aggr( sum({1<OTHER_KEY_PART={$(Only(OTHER_KEY_PART))}, FACT2_YEAR={$(Only(FACT1_YEAR))}>} FACT2_SALES), MY_KEY_PART)

unfortunately for the selected year there are not key values into FACT1 matching FACT2, but that does not interest me, for I need to calcutate sum over all existing key values for selected year into FACT2.

My pivot's dimension is always '-', and I think it's because selection generates an empty-set for aggr dimension.

So I need to find a way to make aggr dimension selection-insensitive.

Any suggestion?

Thanks a lot!

[d_pranskus]

Hi

Would you attache sample application with sample data?

Cheers

Darius

[#3]

Hi Darius!

Here is a sample...

Thanks!

[d_pranskus]

Hi

Just a clue. Could it be that a broblem is related to the fact that when you select Year 2009 and DestIst, it leaves possible values in DescComp with no transactions. It seems that AGGR function and {1} designator works a little bit differently in diemnsions and expressions.

Cheers

Darius

[#5]

nobody has any other ideas?