I am new to the world of big data, I would like to take a question regarding surrogate keys. Considering a distributed data system, creating surrogate keys with hash md5 seems interesting. At the same time, md5 uses attribute concatenation. My question:
Is there a preference for using md5 hash instead of simply using the concatenation of some attributes as the surrogate key?
Great question . I believe both the implementation will work just fine , but from a readablity point of view I think having a set of properties compared to hashed md5 will make sense .