Alexander Farber <alexander.farber@xxxxxxxxx> writes: > 1) If PQconnectdb fails, do I still need to PQfinish the returned pointer? Yes, if you don't want to leak memory. > 2) Similar, if PQprepare fails, do I still need to PQclear its result? Yes, if you don't want to leak memory. > 3) Do I have to PQclear(res) inbetween if I want to prepare another query? They are not comparable actions --- you can do them in either order. In practice though I don't see why you wouldn't PQclear the result of a PREPARE as soon as you'd checked that it wasn't conveying an error. > 4) How do I set the last PQprepare argument, the const Oid *paramTypes? You'd need to look up the OIDs of the parameter types. In practice it's usally a lot easier to write the queries so that the parameter types can be inferred by the backend. regards, tom lane
