Number of dimensions in a dimensional model.
3 posters
Page 1 of 1
Number of dimensions in a dimensional model.
In kimball books, 5-15 is recommended limit for the number of dimensions in a dimensional model to avoid centipede design. Does this number consider role playing dimensions as 1 irrespective of how many roles it plays? For ex. if i have an accumulating snapshot and date has 9 roles, should i consider this as 9 or 1? In my accumulating snapshot after combining dimensions wherever possible i have around 13 unique dimensions without considering various roles. If i include roles, it goes to 30 or so with 9 roles for dates, and a couple of other dimensions being role playing dimensions also. Any comments on this?
arfoot653@gmail.com- Posts : 9
Join date : 2012-02-01
Re: Number of dimensions in a dimensional model.
You would count each role. But also realize the 5-15 suggested range is simply that, a suggestion. It doesn't mean a fact cannot or should not have more than 15 dimensions, but rather is a red flag telling you you need to perform a sanity check. Are these dimensions really necessary or have you over-engineered something?
Re: Number of dimensions in a dimensional model.
I've done more than 30 with success (mostly date dimension FK's).
BoxesAndLines- Posts : 1212
Join date : 2009-02-03
Location : USA
Similar topics
» Recursive Dimensions in Dimensional Model
» Rule based algorithm to convert an ER model to a dimensional model
» Complexities of Relational Model and Simplicities of Dimensional Model
» Fitting number of dimensions
» How to model number of subscribers for a certain point in time
» Rule based algorithm to convert an ER model to a dimensional model
» Complexities of Relational Model and Simplicities of Dimensional Model
» Fitting number of dimensions
» How to model number of subscribers for a certain point in time
Page 1 of 1
Permissions in this forum:
You cannot reply to topics in this forum