On Tue, 2007-02-06 at 14:06 -0500, Merlin Moncure wrote:
> On 2/6/07, Mark Lewis <[EMAIL PROTECTED]> wrote:
> > > actually, I get the stupid award also because RI check to unindexed
> > > column is not possible :) (this haunts deletes, not inserts).
> >
> > Sure it's possible:
> >
> > CREATE TABLE parent (col1 int4);
> > -- insert many millions of rows into parent
> > CREATE TABLE child (col1 int4 REFERENCES parent(col1));
> > -- insert many millions of rows into child, very very slowly.
>
> the database will not allow you to create a RI link out unless the
> parent table has a primary key/unique constraint, which the database
> backs with an index....and you can't even trick it afterwards by
> dropping the constraint.
>
> it's the other direction, when you cascade forwards when you can have
> a problem. this is most common with a delete, but can also happen on
> an update of a table's primary key with child tables referencing it.
>
Hmmm, should check my SQL before hitting send I guess. Well, at least
you no longer have to wear the stupid award, Merlin :)
-- Mark Lewis
---------------------------(end of broadcast)---------------------------
TIP 7: You can help support the PostgreSQL project by donating at
http://www.postgresql.org/about/donate
