Hello,

I have 15 measures and would like to produce reports
which have weeks as the rows and
a) the current years measure
b) the previous year's measure
c) current year's measure less the previous year's measure
as columns.
Some reports would only have only 1 measure
whereas others would have 2 or 3 measures.

Would you define 15 new measures for the previous year's
measures and 15 new measures for the differences?
Is this the best approach?

Thanks
Gary

Thanks
Gary

Gary,

No, as I think you suspect, creating 15 calculated measures for each
real measure when you only want to compare current and previous year
values is not the most efficient way to approach the problem. What you
need is a 'time utility dimension', described in chapter 11 of 'MDX
Solutions' by George Spofford; if you don't have the book, search for
his postings on this ng where he discusses them, eg
http://groups.google.co.uk/groups?hl...com%26rnum%3D1

HTH,

Chris

"GBillins" <gbillins (AT) petvalu (DOT) com> wrote

Can this method be used to include more than 1 measure per
report? I don't understand how I would get for example
get 2 measures M and N in the same report.

M Prior Yr M Diff M N Prior Yr N Diff N
- ---------- ------ - ---------- ------

Thanks
Gary

40dsslab.com&rnum=1&prev=/groups%3Fq%3DSpofford%2Btime%
2Butility%2Bdimension%
2Bgroup:microsoft.public.sqlserver.olap%26hl%3Den% 26lr%3D%
26ie%3DUTF-8%26oe%3DUTF-8%26group%
3Dmicrosoft.public.sqlserver.olap%26selm%
3D3D66285F.F69BE1E5%2540dsslab.com%26rnum%3D1

You can crossjoin a set of measures and a set of "time utility" members, say

select
{ {measures.m,measures.n} * {TU.Actual, TU.Prior, TU.Diff}} on column
from cub

HTH
Bria
www.geocities.com/brianaltmann/olap.html