Re: Prepared statements considered harmful

Peter Eisentraut <peter_e(at)gmx(dot)net> writes:

> - Redefine "prepared" to mean "parsed" rather than "parsed and planned".

Then you would be going very much against the user's expectations.

Driver interfaces expose very clearly to the user an explicit interface to
prepare and execute a query separately. What your proposing is to go behind
the user's back and do what he's gone out of his way to tell you not to do.
You can always choose to prepare your queries immediately before use. Most
drivers even supply an interface to do so in a single step for convenience.

If you've gone to the trouble of saving the prepared query handle you very
much do NOT want the database spontaneously deciding to change the behaviour
of that query (even just the performance behaviour) without warning.

In fact somewhere down the list of my personal wishlist for Postgres is plan
stability which would let the DBA control exactly when plans could change.

--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com