Mathijs Brands wrote:
> On Sat, Mar 24, 2001 at 11:36:56PM -0500, Tom Lane allegedly wrote:
> > Mathijs Brands <[EMAIL PROTECTED]> writes:
> > > How about being able to recompile them (keeping the SQL around in the
> > > system catalogs)? Doesn't Oracle allow you to do something like that?
> >
> > That's another possibility. It's not real clear that there's any
> > advantage to storing rules in preparsed form to begin with --- if
> > we just stored the original text and reparsed it each time it was
> > read, the system would be vastly more flexible, and probably not
> > noticeably slower.
>
> But every bit of performance counts, of course...
The question is if parsing the original query is that more
expensive than converting the printed node tree back into
it's binary representation, what's done now. And then again,
this is only done when the relation is opened and it's
RelationData not found in the relcache. If the relcache
serves well, this happens once per connection.
OTOH, due to toast we don't need to save space in the
pg_rewrite tuples any more. Adding two new attributes to hold
just the backparsed rule qualification and actions
(backparsed from the nodes like done for pg_dump), it might
be easy to create a utility that recompiles rules - for one
or all relations.
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== [EMAIL PROTECTED] #
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [EMAIL PROTECTED] so that your
message can get through to the mailing list cleanly
