Keep a running tally of how many SQL operations have been performed

[fpringle]

Thanks to our dynamic Opaleye effect, we can write an alternative interpreter which, as well as performing SQL operations as before, will also keep a tally of the number of SELECTs, INSERTs etc that have been performed. This is really useful for debugging.

The intended use-case is a sort of bench mark that runs several Opaleye operations for different "sizes", counts the SQL operations, and prints the tallies to the console. This lets us detect if some implementations are ineffecient. For example, for a typical simple read operation e.g. getUserTransactions :: [UserId] -> Eff es [[Transaction]], the number of SELECTs should not depend on the number of Users or their Transactionss. We would expect the number of SELECTs to remain basically constant (O(1)), while the execution time might grow linearly (O(u * t)). If we notice that the number of SELECTs is actually growing linearly, clearly there's an inefficiency in our query design.

I've already used this in a real-world project, and it was remarkably useful to detect inefficient queries.