Re: Effects of cascading references in foreign keys

Bruce Momjian <pgman@candle.pha.pa.us>

From: Bruce Momjian <pgman@candle.pha.pa.us>
To: Bruno Wolff III <bruno@wolff.to>
Cc: Martin Lesser <ml-pgsql@bettercom.de>, pgsql-performance@postgresql.org
Date: 2005-10-29T16:05:33Z
Lists: pgsql-performance
Bruno Wolff III wrote:
> On Sat, Oct 29, 2005 at 13:10:31 +0200,
>   Martin Lesser <ml-pgsql@bettercom.de> wrote:
> > Which effects have UPDATEs on REFERENCEd TABLEs when only columns in the
> > referenced table are updated which are not part of the FOREIGN KEY
> > constraint?
> 
> In 8.1 there is a check to see if the foreign key value has changed and if
> not a trigger isn't queued. In the currently released versions any update
> will fire triggers.
> The check in comment for trigger.c didn't say if this optimization applied
> to both referencing and referenced keys or just one of those.
> If you need to know more you can look at the code at:
> http://developer.postgresql.org/cvsweb.cgi/pgsql/src/backend/commands/
> for trigger.c.

It applies to both.  See
src/backend/utils/adt/ri_triggers.c::RI_FKey_keyequal_upd_pk() and
RI_FKey_keyequal_upd_fk().  The first is for primary keys (pk), the
second for foreign keys (fk). These are called by
src/backend/command/triggers.c::AfterTriggerSaveEvent().  The checks
prevent the trigger from being registered at all if there is no change
in the primary/foreign key relationship.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073