On Wed, Sep 5, 2012 at 3:51 PM, Tom Lane <tgl@xxxxxxxxxxxxx> wrote: > Mike Christensen <mike@xxxxxxxxxxxxx> writes: >> I'm curious under what circumstances Postgres will cache an execution >> plan for a query. > > If you're writing raw SQL, never. The assumption is that the > application knows its usage pattern a lot better than the server does, > and if the application is going to re-execute the same/similar statement > a lot of times, the app ought to make use of a prepared statement for > that. > > Some client-side code (such as the JDBC driver) will make use of > prepared statements under the hood, so a lot depends on context. > But sending plain SQL with PQexec() does not result in any cached plan. Excellent, that's pretty much what I figured (and would expect).. It seems SQL Server and Oracle have some weird caching behavior that's hard to understand and/or predict.. Postgres also seems to be unique in the fact it even has a PREPARE statement.. MS SQL and Oracle only provide that feature through the API.. Mike -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
