Sparse dimensions, how to handle?

View previous topic View next topic Go down

Sparse dimensions, how to handle?

Post  media_agency on Sat May 02, 2009 8:41 am

I've purchased the book "The Microsoft Data Warehouse Toolkit" and
have reviewed it in order to build out a data warehouse solution I am
currently designing. I am stuck in a spot that I don't seem to find
any examples of , or am not able to "make the leap" to correlate it to
any of your scenarios.

My project involves data related to purchasing advertising time or
placements. For the majority of the dimensions they fall right into
the model with a single fact table and multiple qualitative dimensions
related to the fact table.

Where I am struggling is with the invoicing discipline. In this case,
the situation may be that: 1.) no offsetting invoice has been received
for one or more ads, 2.) an invoice has been received but none or some
of the line items may be matched to the ordered ads, 3.) all of the
line items may be matched to ordered ads. The difficulty is that the
current dimensional model covers the ordered ads. The link to the
invoice data is in a single identity field that may or may not be
populated. If it isn't populated, I lack the ability to link to the
invoice set. I do however need to have information about all received
invoices in my cube.

After reading online about this as much as I could find, I believe
this is considered a sparse dimension. On an Oracle platform, they
refer to something called composites to deal with this situation.

So, my questions are:

1. Is this an example of a "sparse dimension"?

2. If so, does the Microsoft platform have anything similar to deal with
sparse dimensions, such as an equivalent to "composites"?

3. What are the "best practices" for modeling this type of dimension?

If I need to provide more information to clarify the scenario, I'd be
happy to do that.

Thanks in advance for any help you can provide. The book has already
proven to be an invaluable resource, so thanks also for providing a
great roadmap.

media_agency

Posts : 1
Join date : 2009-05-02

View user profile

Back to top Go down

View previous topic View next topic Back to top

- Similar topics

 
Permissions in this forum:
You cannot reply to topics in this forum