Hello,

I have in one MDX something like

NonEmpty(
descendants(a.currentmember,leaves) *
descendants(b.currentmember,leaves) *
descendants(c.currentmember,leaves) *
descendants(d.currentmember,leaves),
measures.sales
)

The a and b dimensiopn are quite big (customers and products).

Is the following approach better?

NonEmpty(
NonEmpty(descendants(a.currentmember,leaves) ,measures.sales ) *
NonEmpty(descendants(b.currentmember,leaves) , measures.sales) *
descendants(c.currentmember,leaves) *
descendants(d.currentmember,leaves),
measures.sales
)


Do you have any ideea?

Thanks,
Radu

I expect query optimizer will do NonEmpty on each in second expression, and
then perform inner join of the results.
So, I guess second one is more effective and the order of dimension is also
important.

I also wish more good information on this.

Ohjoo


"Radu Colceriu" <radu_colceriu (AT) hotmail (DOT) com> wrote

After testing both MDX I've got a very strange result:

The second approach uses much less memory (1.6G first approach, 200M
second approach) but there is no speed increase for the second approach.


-Radu



Ohjoo Kwon wrote: