From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Joe Conway <mail(at)joeconway(dot)com> |
Cc: | Darko Prenosil <darko(dot)prenosil(at)finteh(dot)hr>, hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: cross-db queries (was Are we losing momentum?) |
Date: | 2003-04-16 17:35:45 |
Message-ID: | 29147.1050514545@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Joe Conway <mail(at)joeconway(dot)com> writes:
> Darko Prenosil wrote:
>> Unfortunately any other way ends up with first selecting *ALL* records from
>> host ! If there is no such limitation I'll be pretty satisfied with dblink,
>> and will newer ask for "cross-db-queries" again !!!
> Yeah, this is why a proper implementation following the spec is needed.
> If the external access was part of the backend, then the planner could
> be taught to push down qualifiers to the external source where
> appropriate (I think -- maybe Tom will comment on this).
Yes, the newer version of SQL-MED has APIs that allow this sort of thing
to be done. Of course, it's another huge chunk of work beyond basic
SQL-MED ... but at least the roadmap is there, and when we get to the
end of the road we might even find other DBMSes that can speak the same
language. If we invent our own spec the chance of handling cross-DBMS
queries intelligently is nil :-(
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Peter Eisentraut | 2003-04-16 17:41:08 | Re: encoding question |
Previous Message | Tom Lane | 2003-04-16 17:29:01 | Re: Many comments (related to "Are we losing momentum?") |