I have found a trick to fool the system: I use an ORDER BY
response_code 0 ASC LIMIT 1
As unpaid orders receive a response_code > 0, then necessarily the
first record has response_code of 0.
However if more and more orders come into the equation, this means
PgSQL will have to process more records, how is it optimized? I guess
PgSQL will have to find all the records, than order them by
response_code ASC, and then pick the first one.
This is probably not a good practice? Should I go for 2 distinct queries?