Date: Fri, 1 May 2015 11:11:29 -0400
From: Stephen Frost <sfrost@snowman.net>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Andres Freund <andres@anarazel.de>, Peter Geoghegan <pg@heroku.com>,
 hlinnaka <hlinnaka@iki.fi>, Jeff Janes <jeff.janes@gmail.com>,
 Dean Rasheed <dean.a.rasheed@gmail.com>, Bruce Momjian <bruce@momjian.us>,
 Pg Hackers <pgsql-hackers@postgresql.org>
Subject: Re: INSERT ... ON CONFLICT UPDATE/IGNORE 4.0
Message-ID: <20150501151129.GI30322@tamriel.snowman.net>
References: 
 <CAM3SWZTGE0z5_E8wRi9wt3qgOiMpSwtwr5uDcZDGhRZ1_pz6vg@mail.gmail.com>
 <20150501135800.GC22649@awork2.anarazel.de>
 <CA+TgmoYQ185bf=_Z5r0MWRbN2xLgPNMK4N-Q9aAPOC7coZ4HZA@mail.gmail.com>
 <20150501141052.GE22649@awork2.anarazel.de>
 <CA+TgmoZ_X1D_S84aKc-_g6+wU5ri9Q-XYRe6yAMvuwe8wkuwCg@mail.gmail.com>
 <20150501142403.GG22649@awork2.anarazel.de>
 <CA+TgmoYozhG9dcwpVwQ+cpOs+5jakx6MUuXy9zVoXcDSViYqZg@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
 protocol="application/pgp-signature"; boundary="MUgosN8HgFyWu2R6"
Content-Disposition: inline
In-Reply-To: 
 <CA+TgmoYozhG9dcwpVwQ+cpOs+5jakx6MUuXy9zVoXcDSViYqZg@mail.gmail.com>
User-Agent: Mutt/1.5.21 (2010-09-15)
Precedence: bulk
Sender: pgsql-hackers-owner@postgresql.org


--MUgosN8HgFyWu2R6
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

* Robert Haas (robertmhaas@gmail.com) wrote:
> OK.  In that case, I'm a lot less sure what the right decision is.  It
> seems weird for both the BEFORE INSERT and BEFORE UPDATE triggers to
> get a crack at the same tuple, so your way might be better after all.
> But on the other hand, the BEFORE INSERT trigger might have had side
> effects, so we can't just pretend it didn't happen.

I agree that having the before-insert and before-update triggers both
fire is a bit odd, but I also think it's the right thing to do.  If the
statements were independent instead of an INSERT .. ON CONFLICT, then
both sets of before triggers would very clearly fire.

> One idea is to decide that an INSERT with an ON CONFLICT UPDATE
> handler is still an INSERT.  Period.  So the INSERT triggers run, the
> UPDATE triggers don't, and that's it.

Another option would be to have an independent "INSERT-ON-CONFLICT"
before trigger which fires..  but, for my 2c, I'm happy to just fire
both.

I definitely feel that the EXCLUDED tuple should refer to the post
before-insert trigger; having it refer to a tuple that may not have
actually conflicted doesn't seem correct to me.

	Thanks!

		Stephen

--MUgosN8HgFyWu2R6
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQIcBAEBCgAGBQJVQ5ehAAoJEO1sijiDR2RVr7MQAL0+PSmRANqfkvLO5jvIoh6b
yRnlGGzWqteys7bvVcVn3w3eoUaCDRMaMWnbK5FUC1CGA+Q12Gkzsd+fVpWjc54k
8CS3m/9jS7px7owYQV8GJkkb/9+5iZgUYt5spYpwp5HCycE5GVmYc2L0guFPN3qE
qd134xZVwcg7CjFE6SRdEhzoBm6zFMNGgRyNfZzmDFYGepiVY2slDcJAb8XBxnFV
Wt4tCyT4XTzvapEoKlKQKWkkvVyNet27hTN1hAfChQliGJzRAalf4dCqkHlxZ2zT
iBrmbok7+w6Ojf047meOdqs+QvJbh1wmmLd4vIZbBVA63enOtrKbg2yx4WdXwtYE
ONC/u8OBbWQDCbP+RWfR96+WGdkhhW2jDm+E8WLFKcVQLDbA1pUgmyPmp1RiPSoA
pD0QGZPcUpl/oBp4r8EXJ2M1pKzhQe908UdtMbQ5ZoFIm1TYIhYDeBvxyksXxlzn
asQArz46LPTOh3GESQ1XOiOiRDjPFYAl16xu9TEo0fXP83Rz7F5hdWJblPKN14m4
B592V5PLFLcV0UdH49vJ+slJkqhkjyxJ3UVSSsy3UfMujqflNykgQUckjU+yhptL
23A/TuHF0+Rbm19jHiqZDMS9QQsOjjZdSSD86mxtqWbW6L5oGf9eZnlfGWTEA2HY
IzOYC4ejtr/qWooXpjEb
=O9dw
-----END PGP SIGNATURE-----

--MUgosN8HgFyWu2R6--