I find that it is actually faster to do a SELECT from a single record set,
then fetch each individual row and do additional SELECTS to find additional
pieces of data from other databases, rather than to a join between two or
three databases

for example

        SELECT ctryid, dirid, notes from initiative order by postedon limit 3;
        (for each individual record in the above select)...
                select name from country where country.ctryid=iRec.ctryid
                select name from directory where directory.dirid=iRec.dirid

(well this is kind of abstract code but you can follow the logic)

Has anyone else noticed that when you do multiple joins the speed of
PGSQL6.5 goes down dramatically? This is the difference between 2 seconds
for the above and 15 for a join... Is there some way to optimize?


Never retreat. Never surrender. Never cut a deal with a dragon.
_______________________________________________________________
Justin Long                   CIO / Site Editor
616 Station Square Ct         Network for Strategic Missions
Chesapeake, VA 23320          977 Centerville Trnpk CSB 317
[EMAIL PROTECTED]             Va Beach, VA 23463
Check out our site at:        http://www.strategicnetwork.org

Reply via email to