Skip to Content

Dimensional modeling is the modeling approach suggested by Ralph Kimball who advocates building an enterprise data warehouse basis the key facts of the organization and then build the analysis around the same with a strong dimensional model to support various analysis.

What is Dimensional Modeling – let us take an Example. Here our LO extractor VAITM will serve as a good example. I am not taking the VAITM extractor but the fact – Invoice Line Item as an example.

Let us assume a simple Line which basically contains the following facts:

Unit Price
Net Amount
Total Amount

Now let us examine the possible dimensions by which this can be analyzed…..


Analyze sales by Product, Product Variant, Product Brand, Product Category, and Product Type


Analyze sales by Customer, City, Zone, District, and Country. Region


Analyze sales by Day, Week, Month, Quarter, and Year

This would give an idea as to how cube / model design can be approached. Here a particular fact which is Invoice Line Item can be approached. The dimensions are not final and will usually depend on customer requirement and reporting needs but just a simple example to demonstrate how a dimensional model can be arrived at. Also this does not suggest that you keep all the attributes as Navigational Attributes, which I would leave to the architect to decide depending on data volumes and analysis requirements.

This is just my interpretation of Dimensional Modeling for a simple fact but need not conform to pure dimensional modeling views.

To report this post you need to login first.


You must be Logged on to comment or reply to a post.

  1. Former Member
    Thanks Arun for the insight analysis.

    Just want to add that when modelling,there some facts that need to be taken into consideration like the growth of data over time as this will hugely impact on the loading time and space on the database as well as knowing that the datawarehouse will change with time and your design need to cater for that too.

    1. Former Member Post author
      I definitely agree with you … but then issues like sparse dimensions etc do not become evident at the beginning … but then general dimensional design parameters for performance can be followed… I guess this would be common to all cubes in the system….
      1. Former Member
        Deciding how many dimensions you should have and which characteristics are in each dimension is very important. Usually it is best to design the cube for optimal performance and then use a multi-provider to make it more user friendly. There is a good (but detailed) SAP document in and a simple example of cube design on: and
        Regards, Sasha

Leave a Reply