0

I hesitate whether I should add a Country_Dimension or not since I already have a Customer_Dimension which contain some redundant fields such as:

  • continent_name
  • country_name
  • postcode_#
HaveNoDisplayName
  • 8,291
  • 106
  • 37
  • 47
seeusoon
  • 45
  • 6

1 Answers1

0

Those could be two completely different things. A country name present in your Customer_Dimension will probably represent the country of an address your customer is using. Most likely the country he's living (or has lived) in. It can change over time since customers can switch addresses.

A dimension representing Countries will do exactly that, it represents countries. I think you first have to decide what the use of your dimension should be.

Redshark
  • 1
  • 2
  • Thank you for your answer. Let's say I want to analyse the sales per countries, I just have to use the fields in the Customer_Dimension. Even if there is a change in the adress of the customer my Customer_Dimension is going to keep both records, one with the previous adress and one with the new. – seeusoon Nov 13 '15 at 08:59
  • 1
    In a situation like this you don't need to create a new dimension, you're looking at the countries that your customers reside in a the moment the purchase was done. It would be a good idea to alter your analyses to filter out multiple versions of a customers to avoid incorrect calculations (i.e. customers with multiple addresses.) – Redshark Nov 13 '15 at 11:58
  • Hi @Redshark, Thank you for your answer. Yes, I will definitely use slow change dimension type 2 to keep tracks of previous customer address. Then, I think we agree there is no need to create a country dimension. But I still feel weird about that, since in all star schema I saw so far, there is always a location/geographical dimension. – seeusoon Nov 16 '15 at 02:09