I followed the CDVP2 course several months ago and I'm currently busy building a raw data vault.
As primary key for each satellite, we'd like to use the combination of the hash key and the load date.
We are afraid to have duplicates (records with the same hash key and load_date) because of intra-day changes and insufficient load date precision.
In order to make sure that the primary key is unique in each satellite, we are thinking about adding the hash diff into the primary key definition (primary key = hash key + load date + hash diff).
I cannot find any example which uses this primary key structure but would it really be a bad idea?