We do something similar however we also have the ability to delete partitions over a certain age. Hive already has a property inside every table called "retention" that we could consider using.
This code is a good first step but I have one question. Isn't this code rather racey? If we have multiple CLIs running threads they could all be simultaneously deleting tables, and a CLI with a system with a misconfiguration clock could potentially delete all the tables. I think if we do this it should be a stand alone piece.