public inbox for [email protected]  

help / color / mirror / Atom feed

directly construct a query plan
3+ messages / 3 participants
[nested] [flat]

* directly construct a query plan
@ 2004-10-04 17:42 Chengkai Li <[email protected]>
  2004-10-04 19:02 ` Re: directly construct a query plan Tom Lane <[email protected]>
  0 siblings, 1 reply; 3+ messages in thread

From: Chengkai Li @ 2004-10-04 17:42 UTC (permalink / raw)
  To: pgsql-docs

Hello everyone,
I plan to directly construct a query plan without going through the query 
optimizer, by manipulating the various data structures for representing 
query plan tree and parse tree. Is there any facility inside/outside 
postgresql that can be of help? Is there any known issue/difficulty related 
to this?

Thanks!

Steve

_________________________________________________________________
On the road to retirement? Check out MSN Life Events for advice on how to 
get there! http://lifeevents.msn.com/category.aspx?cid=Retirement




^ permalink  raw  reply  [nested|flat] 3+ messages in thread

* Re: directly construct a query plan
  2004-10-04 17:42 directly construct a query plan Chengkai Li <[email protected]>
@ 2004-10-04 19:02 ` Tom Lane <[email protected]>
  0 siblings, 0 replies; 3+ messages in thread

From: Tom Lane @ 2004-10-04 19:02 UTC (permalink / raw)
  To: Chengkai Li <[email protected]>; +Cc: pgsql-docs

"Chengkai Li" <[email protected]> writes:
> I plan to directly construct a query plan without going through the query 
> optimizer, by manipulating the various data structures for representing 
> query plan tree and parse tree. Is there any facility inside/outside 
> postgresql that can be of help? Is there any known issue/difficulty related 
> to this?

Only that your code *will* break in each future release, because we have
no hesitation about whacking those data structures around ...

			regards, tom lane




^ permalink  raw  reply  [nested|flat] 3+ messages in thread

* Re: directly construct a query plan
@ 2004-10-04 22:28 Steve Jackson <[email protected]>
  0 siblings, 0 replies; 3+ messages in thread

From: Steve Jackson @ 2004-10-04 22:28 UTC (permalink / raw)
  To: [email protected]; +Cc: pgsql-docs

Hi Tom,
Thanks!

So if I want to try out some new join algorithm, or even new type of plan 
nodes, before I feel the algorithms/nodes are so good that I want to enable 
the optimizer to consider them as options, is there any better way to 
experiment with possible new plans?

regards,
Steve

> > I plan to directly construct a query plan without going through the 
>query
> > optimizer, by manipulating the various data structures for representing
> > query plan tree and parse tree. Is there any facility inside/outside
> > postgresql that can be of help? Is there any known issue/difficulty 
>related
> > to this?
>
>Only that your code *will* break in each future release, because we have
>no hesitation about whacking those data structures around ...
>
>			regards, tom lane

_________________________________________________________________
On the road to retirement? Check out MSN Life Events for advice on how to 
get there! http://lifeevents.msn.com/category.aspx?cid=Retirement





^ permalink  raw  reply  [nested|flat] 3+ messages in thread

end of thread, other threads:[~2004-10-04 22:28 UTC | newest]

Thread overview: 3+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2004-10-04 17:42 directly construct a query plan Chengkai Li <[email protected]>
2004-10-04 19:02 ` Tom Lane <[email protected]>
2004-10-04 22:28 Re: directly construct a query plan Steve Jackson <[email protected]>

This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox