PG doesn't have bind variables and doesn't cache query plans. The only way to do this currently is to wrap it in a SQL function (which can return rowsets) and call the SQL function.
A more general issue is that you can't cache db_multirows directly in the db/templating universe. I've been thinking about how I might implement an easy way to cache pages built by such queries, for our second development round.
Also ... my guess is that the plan step is only a small part of the time spent building the page. Executing the query and building the multi-row structure is probably taking a lot more time than planning.
