From: | Geoff Winkless <pgsqladmin(at)geoff(dot)dj> |
---|---|
To: | Stephen Frost <sfrost(at)snowman(dot)net>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: INSERT ... ON CONFLICT syntax issues |
Date: | 2015-04-28 14:57:07 |
Message-ID: | CAEzk6fcevLR82Sdr8Cd0Tp9XkLbzKSGyNYng7SM+JaT10pWhjw@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 28 April 2015 at 15:46, Stephen Frost <sfrost(at)snowman(dot)net> wrote:
> +1, NEW/OLD seem pretty natural and I'm not worried about what they look
> like in rules, and their usage in triggers matches up with what they'd
> mean here, I'd think.
>
Since I've stuck my head above the parapet once I figured I'd give m
y 2p's worth:
IMHO
NEW/OLD doesn't fit at all.
In triggers you're applying it to something that (without the trigger)
would be the new or old version of a matching row
, so it's completely intuitive
; in this instance without the ON CONFLICT there would never be a
"
new
"
, because it would be
a
failure
.
MySQL uses VALUES(columnname) to reference the intended INSERT value (what
you might term "NEW") and the target name to reference "OLD". I understand
that people might think the bracketed syntax isn't very pleasant because
that looks like a function, but it seems more reasonable than NEW (can we
use VALUES.columname?); finally I don't see why we need an "OLD" (or
TARGET) at all - am I missing the point?
Geoff
From | Date | Subject | |
---|---|---|---|
Next Message | Geoff Winkless | 2015-04-28 15:12:11 | Re: INSERT ... ON CONFLICT syntax issues |
Previous Message | David Steele | 2015-04-28 14:56:26 | Re: Proposal: knowing detail of config files via SQL |