cancel
Showing results for 
Search instead for 
Did you mean: 

How to design the cube

former_member215567
Participant
0 Kudos

Hi Experts,

Sales data (by Product) is available from the source system. However the analytical reporting is done Brand wise. The association of Product and Brand changes every year. The reports should show the association as on the transaction date. Which is the best way to model the Product-Brand combination?

  1. Product as a dimension member, Brand as a time independent navigational attribute of Product
  2. Brand as a dimension member, Product as a time independent navigational attribute of Brand
  3. Both Product and Brand as dimension members
  4. External hierarchy created on Product with brand as external characteristic.

Please let me know the logic

Thanks

Venkat

Accepted Solutions (0)

Answers (3)

Answers (3)

Former Member
0 Kudos

Hi Venkat,

I would prefer to keep product as a dimension member and brand is time independent characteristics.

Thanks,

Kamal

Former Member
0 Kudos

Hi Venkat,

Brand and Product have 1:n relationship. So, we didn't maintain in one dimension. Because of Cartesian product.

  • Product as a dimension member, Brand as a time independent navigational attribute of Product
  • Brand as a dimension member, Product as a time independent navigational attribute of Brand

In my knowledge we need to maintain Brand and Product as different dimensions.

So you opt the below option

  • Both Product and Brand as dimension members

Thanks,

Purushotham.

former_member182343
Active Contributor
0 Kudos

Dear Venkat,

Does relationship between Brand and product id N : 1.

Means one Brand have multiple products. right ?

The  create 2 dimensions and add product and brand in different dimension.

Otherwise cardinality of dimension increases and leads to performance issues.

Brest wishes,

Vijay