In article <20040203160813.81708.qmail@web13808.mail.yahoo.com>,
Chris Gamache <cgg007@yahoo.com> writes:
> View and rule still apply, tho. Better performance still?
> create temporary table new_money (product text, cents int4);
> create view v_new_money as select product, (cents::numeric/100)::numeric(10,2)
> as dollars from new_money;
> create rule v_new_money_upd as on update to v_new_money do instead update
> new_money set product=new.product, cents=new.dollars * 100 where product =
> old.product;
> create rule v_new_money_del as on delete to v_new_money do instead delete from
> new_money where product = old.product;
> create rule v_new_money_ins as on insert to v_new_money do instead insert into
> new_money (product, cents) values (new.product, new.dollars * 100);
> insert into new_money (product, cents) values ('Flowbee','1995');
> insert into new_money (product, cents) values ('Country Hits','995');
> insert into v_new_money (product, dollars) values ('ThighMaster','39.95');
> update v_new_money set dollars = '14.95' where product='Flowbee';
From where would you get this '14.95'? From your application? If
yes, what type is it? A string or a float? A string would be
inefficient (like SQLs NUMERIC), a float would not be exact (like SQLs
FLOAT).