Improve MDX query in 2005

Hello,
is there way how to improve this MDX query

member [Avg] as [Measures].[Count]/count([Atr1]*[Atr2]),
non_empty_behavior=[Measures].[Count]
select [Atr1].children * [Avg] on 0, [Atr3].children on 1 from MYCUBE

This query compute the average [Count]/[Atr2] of [Atr1] but it is very slow
(for example Atr1 is year and Atr2 is Month and I want the month average of
every year; but in my case Atr1 and Atr2 are not in hiearchy)

I think it can be used FE cache but I don't know how.

Thank you
Ondrej Srubar

This question is wrong. The situation is more complicated. ;(

Ondra Srubar



"Ondra" wrote:

> Hello,
> is there way how to improve this MDX query
>
> member [Avg] as [Measures].[Count]/count([Atr1]*[Atr2]),
> non_empty_behavior=[Measures].[Count]
> select [Atr1].children * [Avg] on 0, [Atr3].children on 1 from MYCUBE
>
> This query compute the average [Count]/[Atr2] of [Atr1] but it is very slow
> (for example Atr1 is year and Atr2 is Month and I want the month average of
> every year; but in my case Atr1 and Atr2 are not in hiearchy)
>
> I think it can be used FE cache but I don't know how.
>
> Thank you
> Ondrej Srubar
>