I'm reposting this from MSDN forums

Hi,

I'm working on a SSAS 2005 project where I migrate an EXCEL pivot table
based on a flat file (coming from an Oracle source). The previous
implementation had most of the members defined twice, first by
description and code and second by code and description, where code is
the key of each customer. In the flat file, for each level with the
need for "dual view support", there are 2 columns, having the same data
but in different format.

So, for example - just for illustration - someone could set the rows to
be

Customer1 (1234)
Customer2 (4553)
Customer3 (3573)
etc

where, someone else

(1234) Customer1
(4553) Customer2
(3573) Customer3

As you might have guessed, this freaked me out, because the flat file
has lots of redundant data and now it is about 120MB fat!

What I'm asking is, how can I implement such a feature on my cube? Is
there a designated SSAS 2005 way to do it elegantly or do I have to
define on my hierarchies each level that requires "dual view support"
twice?

I presume you have a customer dimension in SSAS.
Do you load your flat file into a database?
if yes, you can create a calculated column in your DSV (data source view) to
create your description (id) column and the (id) description column. In this
case you just need to have the description and the ID columns in your
database.

if you use the flat file directly in SSAS without storing the results in a
database, then you have to keep the labels in this file because you can't
create a calculated column (which is an SQL syntax)

In your dimension you can create 2 attributes one based on the description
(id) column and a second based on the (id) description column. The users can
now choose which one they want.



"KelMan" <listmail_nospm_ (AT) intrip (DOT) gr> wrote