>Hi all,
>
>I'm having a discussion with my boss and a fellow co-worker who is a dba about the usage of surrogate keys for primary keys in tables (he calls them "artificial keys").
>
>My colleagues on the other hand are saying that it's a mistake to use "artificial" keys because they aren't self documenting in the code. They

I had a similar argument with someone else about all of this. They used the same argument about it being more self-documenting (ie. showing the implied relationships between fields and their uniqueness). I just don't buy it. I'm maintaining a system right now that uses composite keys and it's a nightmare. Fields that are supposed to be unique aren't always, which means sometimes the routines based on those tables return odd results.

I'd stick with surrogate keys and just create indexes on the would-be composite keys if you want to show some sort of relationship between fields. I think it's just a much safer way to go.