I am currently using Microsoft Analysis Services 8.0 . When I'm designing
storage and aggregating my cube , if i use more than 20 Aggregation in the
cube I get incorrect data.However if I select less aggregations or If i use
ROLAP storage ,everything comes out OK. I seem to be only having problems
with HOLAP and MOLAP .The data is less than its suppose to be. This is a real
problem ,because accessing cubes with little aggregations is really slow.
Please help !

Aggregations should not have any effect on cube's results.

My guess is that some other factor is involved and you blame the
aggregations design instead.

Missing data in cube is usually caused by missing dimension members.

Did you process the cube's dimensions prior to cubes process in both cases?

Did you process exactly the same data slice in both cases?

You can check missing joins to (missing)dimension members by optimizing the
cubes schema (in cube editor look in the tools menu).
Then each dimension's join that had been optimized will be sensitive to
missing dimension keys and will alert you in the processing window. For other
dimensions
you should check manually that all dimension keys in facts exist in those
dimensions as well.


"Marrow" wrote:

Check for column names in the fact table and the dimension tables that,
accidently have the same names, and create joins that are not correct. You
can see this in the cube editor. Be sceptic og multiple joins between one
dimension table and the fact table.

Regards
/Thomas Ivarsson


"Marrow" <Marrow (AT) discussions (DOT) microsoft.com> wrote