Here's the issue...

We're running SQL Server 2005 w/ all services.
We have a table called 'PartProduct' which has 3 columns: PartNumber,
PartQty, and Product. The combination of PartNumber+Product is the key;
a PartNumber can appear on many products, and a Product has many
PartNumbers.

I can create a cube which shows me a pivottable with PartNumbers down
the Y axis and Products across the X axis, so that I can compare how
many of each Part is used on each Product. What I NEED, though, is a
comparison how many Parts each Product has in common, whether by the
number of different parts or the sum of different parts:

Source Table:
PartNo PartQty Product
------ ------- -------
PartA 4.0 ProductA
PartA 3.0 ProductB
PartA 4.0 ProductC
PartB 1.0 ProductA
PartB 1.0 ProductB
PartB 1.0 ProductC
PartC .5 ProductB
PartC .5 ProductA

PivotTable from a cube I've created:
PartNo ProductA ProductB ProductC
------ -------- -------- --------
PartA 4.0 3.0 4.0
PartB 1.0 1.0 1.0
PartC .5 .5

What I need (#1 - total parts used between a pair of products)
(Note that where a product crosses itself, the number of parts is
doubled, since it's a pair of products):

Product ProductA ProductB ProductC
------- -------- -------- --------
ProductA 11.0 10.0 10.5
ProductB 10.0 9.0 9.5
ProductC 10.5 9.5 10.0

What I need (#2 - Number of different PartNos in a pair of products):

Product ProductA ProductB ProductC
------- -------- -------- --------
ProductA 3 3 2
ProductB 3 3 2
ProductC 2 2 2

I need a few other stats, but I think if someone can help me figure out
how to generate these two tables, I can figure out the rest.

Any suggestions are DEEPLY appreciated. I'm an Analysis Services
neophyte and I'm learning as fast as I can.

Thanks, all!

- Tom