How do we model data from multiple systems when they store copies of the same business concept (Customer, Person, Organisation, Product etc), using a Data Vault modeling approach.

 TBD

This pattern should be applied when you cannot guarantee all source applications are using the same shared id’s for concepts. 

There are a number of examples where the use of the same id in different context across different source applications can cause issues. 

For example the id of “A1” for the “ABC123” organisation in the CRM source applications is the different to the id of “123A” for the “ABC123” supplier in the FMIS source application.   

[add data view]

Another example is when the id of “A1” identifies the “ABC123” organisation in the CRM source application, but the same id “A1” identifies a different organisation for example “ANT167” in the FMIS source application.

[add data view]

To mitigate these issue we create a collision code, usually a code relating to the source application, for example “crm” and “fin” which is  concatenated with the id to make a unique key, for example “crm_A1” and “fin_A1”

[add data view]

AGAINST

• Golden Record – Data Vault > Combined Hub
• Golden Record – Data Vault > Unique Hubs > Combined Hub