[bcrosby95]

But MySQL's equivalent, insert...on duplicate, does not cause extra auto increments. IMHO postgres' is a much larger problem.

That said, I say that in my experience based upon locking selects in MySQL - when the row doesn't exist is a big problem, oftentimes leading to deadlocks. So insert... on duplicate key is a life saver. I don't have as much experience using postgres, so I'm not sure how much of a problem that is there, so it's possible the "on conflict" isn't as useful.