Date: Fri, 3 Nov 2017 09:35:39 -0700
From: Peter Geoghegan <pg@bowt.ie>
To: Simon Riggs <simon@2ndquadrant.com>
Cc: Robert Haas <robertmhaas@gmail.com>, Stephen Frost <sfrost@snowman.net>,
	Michael Paquier <michael.paquier@gmail.com>,
	PostgreSQL-development <pgsql-hackers@postgresql.org>
Subject: Re: MERGE SQL Statement for PG11
Message-ID: <20171103163539.GA17026@marmot>
References: 
 <CANP8+j+FPEqpd8O3scwOKUgRvtWB5r5RhcCuJ8nn6VU_wc2Nng@mail.gmail.com>
 <CAH2-Wz=AxEKE7sJEukPN3Vgu6sU5QwuLCBHuF-oJbkOLWsxTQw@mail.gmail.com>
 <CANP8+j+FM8ep1Cprbjaeknzq1pRAaxTMvX=hYsc4eGcj13fz7w@mail.gmail.com>
 <CA+TgmoYbytQT1pZDOK6t-XD=jWffDXaHoubsU-_mqJU6W0oK0w@mail.gmail.com>
 <20171030185511.GJ4628@tamriel.snowman.net>
 <CANP8+jJopkeYT_h_u=jb6euKVnXqRc77CiOk+hSwxOC-xtP0kQ@mail.gmail.com>
 <CA+TgmoYKesHek379ny4HtAi-nGszj_2VrQ4uvU60qV3BNSkGFQ@mail.gmail.com>
 <CANP8+j+iAAvW1VqTBkW7qbq+vB8GD+xYmgQoL5tvrhO0dBcXww@mail.gmail.com>
 <CA+TgmoYOyX4nyu9mbMdYTLzT9X-1RptxaTKSQfbSdpVGXgeAJQ@mail.gmail.com>
 <CANP8+j+JxXs8RWWpz2yRZdbWMti6M5rxyKZDY-yLvDt3ROOYMA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Disposition: inline
In-Reply-To: 
 <CANP8+j+JxXs8RWWpz2yRZdbWMti6M5rxyKZDY-yLvDt3ROOYMA@mail.gmail.com>
User-Agent: Mutt/1.5.24 (2015-08-30)
Precedence: bulk
Sender: pgsql-hackers-owner@postgresql.org

Simon Riggs <simon@2ndquadrant.com> wrote:
>>> The *only* behavioural difference I have proposed would be the *lack*
>>> of an ERROR in (some) concurrent cases.
>>
>> I think that's a big difference.  Error vs. non-error is a big deal by
>> itself;
>
>Are you saying avoiding an ERROR is a bad or good thing?

Are you really asking Robert to repeat what has already been said about
a dozen different ways?

That's *not* the only difference. You need to see a couple of steps
ahead to see further differences, as the real dilemma comes when you
have to reconcile having provided the UPSERT-guarantees with cases that
that doesn't map on to (which can happen in a number of different ways).

I don't understand why you'll talk about just about anything but that.
This is a high-level concern about the overarching design. Do you really
not understand the concern at this point?

>> also, the non-error case involves departing from MVCC
>> semantics just as INSERT .. ON CONFLICT UPDATE does.
>
>Meaning what exactly? What situation occurs that a user would be concerned with?
>
>Please describe exactly what you mean so we get it clear.
>
>The concurrent behaviour for MERGE is allowed to be
>implementation-specific, so we can define it any way we want.

Agreed -- we can. It isn't controversial at all to say that the SQL
standard has nothing to say on this question. The problem is that the
semantics you argue for are ill-defined, and seem to create more
problems than they solve. Why keep bringing up the SQL standard?

-- 
Peter Geoghegan


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers