Yes, I thought you would find that to be your next problem.

Sorry, too tired now to think of a solution that does not involve two queries. It would all depend on the data. How much there is, how dynamic it is. But I think with the right indexes, given the small number of columns in your schema, you could fetch any existing data very quickly if you know a unique user ID. If you can do an efficient check/fetch, then your task of deciding what to pass in the insert/update statement becomes much simpler.