Let’s say I’m building a Person hub and I get a mix of ids from sources. Some of the IDs are source-specific (i.e. meaningful only within that source). Others are meaningful across sources (e.g. drivers license number, government id number). For example:
I think the following are correct, with respect to Data Vault modeling:
- All of the ids should be in a single Person hub
- The source-specific ids should use a BKCC to keep the ids from mixing across sources
For the cross-source ids (e.g. drivers license, government id number) I see two options:
- Use a BKCC for each type of cross-source id (e.g. “DL”, “GOVID”), being careful not to overlap any source BKCC identifiers with these BKCCs.
- Use a concatenated business key, e.g. “GOVID-123456” with no BKCC
Are these reasonable options? Is option 1 within the “spirit” of a BKCC? Are there better options?
Thanks for guiding a Data Vault newbie
Keep in mind mate, and as I have said before, BKCC is not source driven. Only use BKCC sparingly when there could be a clash. The moment you default BKCC to source then you will be building a convoluted mess, source-system vault, ala fake vault.
Never concatenate business keys, you’ve modified data in this regard
@patrickcuba Thanks for the reply. I do understand that it’s not appropriate to default a BKCC to a source as a general rule, as then you forego the ability to passively integrate when sources share a common business key. And I have read your Rose article
In my case, it is possible for XYZ’s keys to overlap with ABCs as these are separate systems generating the IDs. I don’t see how it’s possible to have anything but a source-system id for these ids.
For the Driver’s License and Government ID I do want them to integrate as they are the same across sources. However a Driver’s License and Government ID could in theory overlap, as I don’t have any guarantee as to their formats.
The general advice on government issued ids is that they are not used as business keys by legislation, at least in Australia this action is not allowed and are anyway something that could be considered PII and should be locked away in a PII satellite.
For reference: Chapter 9: APP 9 — Adoption, use or disclosure of government related identifiers - Home