I read in the Book of OHDSI that the “ETL Flow” is generally to create the person table first followed by observation_period, dimensional tables (provider, care_site, etc.) then visit and clinical event tables.
However, the person table has a provider_id and location_id field, so wouldn’t you need to create those first? I’m using rabbit in a hat to document the ETL and was starting with person but am unsure what to fill in for provider_id given that it is not sourced for the source database, but is contingent upon the cdm.provider table being created.