Kimball Forum
Would you like to react to this message? Create an account in a few clicks or log in to continue.

Question On Conformed Dimension design

Go down

Question On Conformed Dimension design Empty Question On Conformed Dimension design

Post  mru22 Fri Aug 19, 2011 10:47 am

After reading Kimball's section on Conformed dimensions and making some changes in my project, I have question related to design.

I have a "Person" dimension that is used in 4 different fact tables. I am creating a fifth fact table for Claim Payment information where there is a claim associated to each claim payment.

The OLTP database structure looks like the following

Person One to Many Claims One to many ClaimPayments

When I bring the data into the warehouse, I have a key directly from Person to claim payment And then Claim directly to claim payment.

By adding the PersonName and Social Security Number to the Claim table I can know which person exists without having to snowflake.

My question is, is the above duplication of columns in two different dimensions usually the way conformed dimensions are used and are they preferred over an outrigger that leads to small snowflaking ? I can see many of my dimensions having duplicate data in order to avoid snowflaking I have kept the column names identical but still not sure if this is the correct way for not joining dimensions but being able to identify relationships between them without fact table joins.


Posts : 34
Join date : 2011-06-14

Back to top Go down

Back to top

- Similar topics

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