The concern I have though with the mapping as it stands today is the passing of the auto_increment fragment of the sql-type to the envers table. I honestly don't see the value in passing this part of the sql-type string, so perhaps that needs to be filtered as well.