Re: [EXTERNAL]: Re: [EXTERNAL]: Re: UPSERT in Postgres

Am Mon, Apr 10, 2023 at 01:33:41PM +0200 schrieb Francisco Olarte:

> > > > This the part that's always eluded me: How does the client, the
> > > > UPSERTer, come to hold an id and not know whether or not it's already in
> > > > the database.
> > >
> > > This is extremely easy to do if you have natural instead of surrogate keys.
> > >
> > > I work in telephony, upserting the last incoming call timestamp for a
> > > phone number will be exactly that.
> >
> > timezones ?
> > DST ?
>
> A timestamp is a point in the time line, this is what I insert, just a
> real number marking a line, timezones and dst are presentation stuff.

Indeed, as is the assumption which time line the numbers are
referring to. Hence the incoming call timestamp is usable as
a (natural) PK with respect to a given time line only, right?

> > spoofing ?
>
> ¿ Of what ?

The time stamp. But then I assume that is obtained on the
logging system.

All I really wanted to hint at is that "incoming call
timestamp" may work pretty well in given settings but does
not _always_ make for a "unique enough" key.

Karsten
--
GPG 40BE 5B0E C98E 1713 AFA6 5BC0 3BEA AC80 7D4F C89B