(Newbie) Outer Join Simulation

I have two tables invoice(fact table) and agents(dimension). Not all invoices have an agent. How do I use the agents table as an outer join? When I include agents as a dimension and look at all agents, I only get invoices that have an agent. I know I could create a view in SQL but I want to do this in the cube.

Thanks,

Bill

You can create an unknown member in your agent dimension table and connect all fact records without an agent key to that member. You will have to use TSQL to this.

SSAS2005 can help you with doing this without writing code. Have a look at inferred members in BOL!

Regards

Thomas Ivarsson

"Missing" Data / Data Not Making It Into AS2005 Cube

All of the data in one of my fact tables in the relational db doesn't seem to be making into the cube. However, all of the data from our sales fact table is indeed making into the cube.

There is no filtering going on in the Named Query in the DSV that brings in this data. What could it be? The table's relationship? Dimension Usage?

I also developed a small test cube and was able to see all of the data. I would hate the thought of having to rebuild/redevelop the cube from scratch.

I've tried everything that I could think of.

Any thoughts?

Hello. One possible explanation can be that the service account, running SSAS2005, do not have permissions in your relational db. Are all you fact tables in the same database /data warehouse? Do you have several data source systems?

Regards

Thomas Ivarsson

|||I think I'm making progress on the troubleshooting.

I just deleted a lot of the dimensions in the Dimension Usage tab for this particular measure group. In doing so, all of the data is now there.

I will slowly add the dimension relationships back to see which one is the cause and then will try to understand why that is the case.|||I figured it out.

# of files in group, CPU, and RAID

Not sure if this is the right forum...
I have a fact table that ha sits own filegroup on a dedicated RAID 10 array.
Somewhere in Books Online I read that you should create a file in the
filegroup for each CPU in the server. My understanding is that in SQL 2000,
this triggered the server to spawn multiple threads to handle the I/O. My
questions are:
1 - Does SQL 2005 need multiple files to spawn multiple threads for I/O?
2 - If not, being on a RAID 10 array, does it help in any way to have
multiple files in a group?
Thanks,
-- Brian"Brian Shelden" <BrianShelden@.discussions.microsoft.com> wrote in message
news:236B2CF8-F24E-43B8-A3EF-0F7D41CEE486@.microsoft.com...
> Not sure if this is the right forum...
> I have a fact table that ha sits own filegroup on a dedicated RAID 10
> array.
> Somewhere in Books Online I read that you should create a file in the
> filegroup for each CPU in the server. My understanding is that in SQL
> 2000,
> this triggered the server to spawn multiple threads to handle the I/O. My
> questions are:
> 1 - Does SQL 2005 need multiple files to spawn multiple threads for I/O?
> 2 - If not, being on a RAID 10 array, does it help in any way to have
> multiple files in a group?
> Thanks,
>
Multipost. Answered in .server.
David