Hello all Analysis Server users out there!

I am using SSAS 2005 with SP1 installed, and I have got a question
concerning the setup of several dimensions derived from one table.
What I mean to say is: if I got several columns in a table and want a
dimension on each one of them, can I build a dimension on the id column
(where all the others depend upon) and add the columns to that dimension as
attribute?

I find it complicated to explain precisly what I mean. I added an example
here, sorry it got so long:


---Example---
Say the table is your standard order table, with following columns: (This is
simplified, I left out the not important columns, and denormanlized it on
customername for this example.)
idOrder (int)

ordernumber (varchar)
customername (varchar)

order date (datetime)
billing date (datetime)
shipping date (datetime)

PayCash (int, yet contains only 0 and 1 as flag)
PayCreditCard (same as above)

ShipUPS (same as above)
ShipPostal (same as above)

Now, I want to build dimensions on all these.

The Date dimensions are obvious to be useful. After I build them using the
wizard, I notice that they all boil down to the id of the table as their key
(which is correct of course!)

The varchar field seem not to be a good candidate, as they will be of the
same magnitude of the table itself. I need them later for drilldown
scenarios, so I need to include them in the cube. And, I notice again that
this is really a dimension of the id column and the varchars are just
attributes of that.

Finally, the integer flags are used to group the orders. I need a dimension
on each of those as well. I again notice that really these are only
dependent and attributes of the id column again.

So, there you have 9 fields that all depend on the id column. I made own
dimension of all of them, yet I wonder if it is not better to have one
dimension which contains all of these as attributes?!

Thanks for reading!
Ralf