Is using the Cube wizard in SSAS where it is decided to use UDM - or where?

Can't really find any tutorials on using UDM - just many discussions about
how great it is if/when it can be used.

Don't qujite understand your question.

In SQL BOL, under SQL Server Analysis Service, you can find "Unified
Dimensional Model". Does that help? If not, please let us know what we
can do to improve it.

There's also an MSDN article by Paul Sanders talking about UDM here:

http://msdn.microsoft.com/library/de...IntrotoUDM.asp

HTH,
Chu


Joe wrote:

Yes - I have seen these articles. My question goes how to implement it.

The adventure works tutorials have you point to AdventureWorksDW - where the
Fact and Dimension tables are already created. Isn't the purpose of UDM to
create Fact and Dimension tables via a connection with an OLTP (or other)
system?

So my question is "how can I practice say using AdventureWorks OLTP DB and
by using UDM methodology create a cube?"

"Chu Xu (MSFT)" <iloveseahawks (AT) gmail (DOT) com> wrote

SSAS is not designed to be used in front of an OLTP database, but is more
powerfull against datamarts and datawarehouse databases.
Like the adventure works sample.
You have the OLTP database, the SSIS packages to convert this database into
datamarts (DW database) and finally there is the SSAS cubes.

If your source database use a simple design, maybe you can create views on
it and then use these views in SSAS to create your cubes.
This option is good for realtime scenario too.
But due to the impact on the source DB system, its not a recommanded option.

Using a DW or Datamart allow you to have a better control on what you do
with your data.
You can transform data, synchronize multiple sources, do some cleansing and
isolate the access from the OLTP database, so the impact for the end user is
small.
Standard transformations are: creating slow changing dimensions; adding
business rules; grouping members (like creating some groups of products for
reporting purpose not existing in the source OLTP)...



"Joe" <hortoristic@gmail dot com> wrote

I still really appreciate everyones comments - but I'm still getting reasons
why NOT to use UDM - not how to use it or implement it. I understand using
the typical data mart is still best sometimes but still want to try UDM.

"Jeje" <willgart (AT) hotmail (DOT) com> wrote

So just to clarify - When I create a data source directly to an OLTP system,
use SSAS to step through the Cube wizard - that is what UDM is? Whereas the
traditional way my data source is a data mart?

I practiced the above on a Client and Client address table from the OLTP
system and the cube wizard created a dimension and fact table inside my SSAS
project - is this what UDM is?

"Joe" <hortoristic@gmail dot com> wrote