On Fri, Jan 31, 2003 at 11:21:05PM -0500, Tom Lane wrote:
> will trillich <will(at)serensoft(dot)com> writes:
> > is this a stupid idea?
> > [create view such that]
> > then we can do things like
> > ->execute("select parent_table from $table")
> Uh, do you really want to impose overhead on *every* query to make
> that one operation faster?
that's what i was asking. is that a lot of overhead?
> I'd go with a query on the system catalogs, for example
> regression=# create table p (f1 int);
> CREATE TABLE
> regression=# create table c (f2 int) inherits(p);
> CREATE TABLE
> regression=# select relname from pg_class where
> regression-# oid = (select inhparent from pg_inherits where
> regression(# inhrelid = (select oid from pg_class where relname = 'c'));
my bad, i probably used the wrong terminology.
this isn't inheritance -- it's many-to-one relationship between
tables. below, org can have many projects; projects are tied to
people via many team records.
org.id => many projects to one org
person.id => many teams (projects) for one person
project.id => many teams (people) for one project
from what i can tell, the trigger function calls are all there
is to indicate the use of foreign keys. and there's some serious
hair in parsing those \0 strings to get what you want out of
them. seems like there has to be some other mechanism that'd
work with less obtusification.
There are 10 kinds of people:
ones that get binary, and ones that don't.
http://sourceforge.net/projects/newbiedoc -- we need your brain!
http://www.dontUthink.com/ -- your brain needs us!
Looking for a firewall? Do you think smoothwall sucks? You're
probably right... Try the folks at http://clarkconnect.org/ !
In response to
pgsql-general by date
|Next:||From: Andrew McMillan||Date: 2003-02-01 04:43:48|
|Subject: Re: [NOVICE] For each record in SELECT|
|Previous:||From: will trillich||Date: 2003-02-01 04:28:24|
|Subject: paradigm sanity check needed [long]|