From: | Decibel! <decibel(at)decibel(dot)org> |
---|---|
To: | Stephen Greensmith <s(dot)greensmith(at)ncrl(dot)co(dot)uk>, pgsql-novice(at)postgresql(dot)org |
Subject: | Re: Triggers and Function. |
Date: | 2008-08-27 16:34:24 |
Message-ID: | 20080827163424.GJ93414@decibel.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-novice |
Adding -novice back in.
On Tue, Aug 26, 2008 at 08:34:45AM +0100, Stephen Greensmith wrote:
> Thanks its looking that way.
>
> Why because I need to know if the application is setting a value in a
> column or not. A bit like having a default on an update as well as an
> insert.
Ok, but why do you need to know if the field appears in the UPDATE, even
if it's value doesn't change?
> On Fri, 2008-08-22 at 06:38, Decibel! wrote:
> > On Aug 18, 2008, at 6:01 AM, Stephen Greensmith wrote:
> > > How can I identify if a column is actually part of the UPDATE that
> > > fired
> > > the trigger.
> > >
> > > For example
> > >
> > > If I have a table with three columns col1, col2 and col3. The types
> > > don't matter to the question but may to the answer.
> > >
> > > I run "UPDATE example set col1 = 'NewVal', col2 = 2 where col1 =
> > > 'OldVal';"
> > >
> > > A BEFORE TRIGGER can test OLD.* against NEW.* and see col1 has changed
> > > col2 may or may not, OK to track what has actually changed, but how
> > > can
> > > I find out that col3 was not part of the update?
> > >
> > > I need to know if the client has updated a column, even if it still
> > > has
> > > the same value.
> >
> >
> > I'm pretty sure this isn't possible. Why do you need it?
--
Decibel!, aka Jim C. Nasby, Database Architect decibel(at)decibel(dot)org
Give your computer some brain candy! www.distributed.net Team #1828
From | Date | Subject | |
---|---|---|---|
Next Message | Bruce Hyatt | 2008-08-27 20:37:17 | Re: Triggers and Function. |
Previous Message | Steve T | 2008-08-26 07:57:41 | Dump/Restore compatibility |