Hi,

We have a scenario to handle 5 distinct count measures and
additional regular metrics. We have designed 6 different cubes, on
for each distinct count measure and one for all other regula
metrics. On top of these cubes, we designed a virtual cube t
include all the measures

We have around 20 Million records in the fact table and 15 dimensions
We are successful in building and processing the cubes with 0
aggregations for distinct count cubes and 10% aggregations for th
regular cube. However, when accessing the reports through MS Offic
Web Components, the reports performing very very slow, sometime
even up to 20 minutes to open the report

I tried setting up 5% aggregations for distinct count cubes eve
changing the aggregation level to top level only, but neve
succeeded. Could you please share your experiences whether anybod
handled this kind of scenario. Unfortunately we don't have 64 bi
environment. Whether it is feasible to handle this situation at al
with OLAP cube

Saga

you have to use the usage based optimization to add required aggregations in
your cubes.

distinct count measure and dynamic distinct count measures needs 2
differents aggregations design!
DCount measure need standard aggregations design while dynamic dcount
measures (or calculated measure) need specific aggregation design (which is
at the customer level when counting customers).


"sagar_goli" <sagar_goli (AT) yahoo-dot-com (DOT) no-spam.invalid> wrote

I might be wrong, but I think that with distinct count creating aggregations
is not relevant.

"Jéjé" wrote: