Hi,

I am looking into switching to SQL Server 2005 and have problem
understanding how better migrate application.
I have a data warehouse that is about 300GB. I have all data in SQL Server
tables and all data there is at lowest level (client/date/etc). I have over
3mln clients, so my choice is not to have client dimension. So my Analysis
Services DB contains just dimensions that are based on client data (start
age, current age, gender, etc). When during analysis I need to get a list of
actual clients (lets say list of ClientKey + few other attributes) , I can
perform drillthrough.
Also I choose not to include client dimension because such a huge dimension
would definately slow down AS querying.
I am studying AS 2005 and now with new attribute logic I cannot seem to
understand how would I be able to reproduce diferent granularity in SQL
Server and AS. Also, from what I understood drillthrough is also redesigned,
so now all fields that are showed in the drillthrough should be included in
cube as attribues or dimensions.
So if I have a requirement that I will have to see ClientKey in
drillthrough, Does that mean that I will have to introduce Client as
dimension?
How well AS will handle dimension with 3-10mln records (depends on company
that buys our product)?
Are there any other options for me?

Thanks in advance.

Some of the concerns regarding drillthrough in AS 2005 are discussed in
an earlier thread of this newsgroup:

http://groups.google.com/group/micro...olap/browse_fr
m/thread/b67c7fc032e313a3/c690266b4ca51834?lnk=st&q=Limitations+of+AS200
5&rnum=1#c690266b4ca51834

There is also this MSDN paper describing drillthrough:

http://msdn.microsoft.com/library/de.../en-us/dnsql90
/html/sql2k5_anservdrill.asp

T.K. Anand
Microsoft Corporation

July 2005

Applies to:
SQL Server 2005 Analysis Services

Summary: Discover the new Analysis Services 2005 drillthrough
architecture. See how to set up drillthrough in Analysis Services 2005
and get guidance on migrating drillthrough settings from Analysis
Services 2000 databases. (10 printed pages)
...

- Deepak

Deepak Puri
Microsoft MVP - SQL Server

*** Sent via Developersdex http://www.developersdex.com ***

Hi,

The AS2005 can considerably better handle huge dimensions. I have introduced
some such flat dimensions (order, invoice) with more then 5 mln members.
With AS2000 it was inpossible (unendurably) because of bad performance.

Vladimir Chtepa

"Nobody" <Nobody (AT) nobody (DOT) com> wrote