Re: Very strange performance decrease when reusing a PreparedStatement

On Sun, 2009-05-03 at 15:08 -0400, Roland Roberts wrote:

> At least PostgreSQL has the ability to tell the parser to reparse
> every time you ask it to prepare the statement. Oracle doesn't. The
> solution for Oracle isn't relevant, but the real point here is that
> your code may have to become smarter to help out the optimizer. We
> reparse on every execution, but can't control Oracle's server-side
> cache. PostgreSQL lets you control that with prepareThreshold=0. If
> we'd had that option with Oracle, we would have had an easy solution.
> I'd say PostgreSQL's implementation for this case is a good one.

We have a good feature there, definitely. Although there is still more
to be done.

Oracle's cursor_sharing features are slightly ahead of Postgres,
specifically around the "SIMILAR" option.

We talked a few years back about an "Option node" that would select
sub-plans at runtime based upon the actual values of the bind variable.
Nobody's got around to doing anything with that yet.

--
Simon Riggs www.2ndQuadrant.com
PostgreSQL Training, Services and Support