Using natural keys is what actually can prevent duplicate rows. In your above example, if email is the PK, there would be no duplicates. But adding an id as a P

Using natural keys is what actually can prevent duplicate rows. In your above example, if email is the PK, there would be no duplicates. But adding an id as a PK would essentially keep your database with duplicates: (1, 'bob', 'bob@bobco.com') (2, 'bob', 'bob@bobco.com')