Re: Surrogate Keys: an Implementation Issue

paul c wrote:

> > Name (pk) City
> > ===== ======
> > Palu New York // Lastweek
> > Palu New Jersey // yesterday
> > Paul New Jersey // today
> >
> > Since my primary key has changed, I need some immutable identifier so I
> > know these are in fact part of the same set.
> >
> > Not sure how a timestamp helps here.
> >
>
> 'Lastweek', 'yesterday', 'today' look like timestamps to me!
>

Sure, but how does the DBMS know that these three rows are somehow related? How would it even produce the result? That is what I'm getting at, and a timestamp does not help. Received on Thu Jul 20 2006 - 22:25:59 CEST