I've been tooling around with the Partition Aggregation Utility and have a
few questions on how to actually set up an aggregation.

The cube has three dimensions with the following levels:

Date
-Year
--Month
---Day

New Account
-YesNoIndicator

Sales Organization
-Sales Group
--Account Manager
---Sales Account


For the most part, end users will be querying based on the Date and Sales
Organization dimensions. The most common report we have will be Day/Sales
Account based on a date range and Account Manager.

My question is: what aggregation do I set up for this?
--Date: Day
--New Account: All
--Sales Organization: Sales Account

The "all" selection has me confused as to what it actually does. Does this
mean it just ignores that dimension for aggregating? Can some explain?

Also, we have partitions which 100 million rows in it. If I set the row
count equal to one, will this mess up with the aggregations I set, as long
as they're not performance based? My reluctantcy in setting them equal
(Manage Partition Count) to an estimate based on the row count is that it
takes forever to actually edit that partition due to the large number of
records.

Thanks in advance!

ALL. This does in-effect remove the dimension from the aggregation.

There is an excellent discussion of all of this in the following two
downloads:
1) AS Performance Guide
http://www.microsoft.com/technet/pro.../ansvcspg.mspx
2) a webcast done by Len Wyatt from Microsoft which discusses the
aggregation portions of the white paper.
http://www.microsoft.com/downloads/d...DisplayLang=en

usage-based optimization wizard). As you are manually creating aggregations,
they will have no effect if you set them to 1.

Hope you find these useful.
--
Dave Wickert [MSFT]
dwickert (AT) online (DOT) microsoft.com
Program Manager
BI SystemsTeam
SQL BI Product Unit (Analysis Services)
--
This posting is provided "AS IS" with no warranties, and confers no rights.


"Jesse O" <jesperzz (AT) hotmail (DOT) com> wrote