From what I am reading, I think it depends on the context of the systems you have, the information you need and how fast you need it. It looks like some folks prefer not to use surrogate keys because of storage and performance issues. One article talks about a passenger key that references the dimension table. Try reading https://www.sciencedirect.com/topics/computer-science/dimension-table
Also please elaborate on the context of your question.
[login to unmask email]
Freelance Production Assistant
Freelance Data, Technology and Science Writer
Many natural keys are already unique and complete. If they are, I'd prefer to use them instead of a surrogate key.