MIME-Version: 1.0
In-Reply-To: <20140423105630.GQ10046@momjian.us>
References: <29030.1396993582@sss.pgh.pa.us>
 <CAM3SWZQzOxmryVe_bdu9WhqbJTBoiqM7knvKm0oRYxvfsFnz7Q@mail.gmail.com>
 <534475B7.6020908@dunslane.net> <5344EAA4.1050605@vmware.com>
 <CA+TgmoZucoe-w5P_16A-gdfRBFCCu7PD17YO-KqrYPBhb5TxAA@mail.gmail.com>
 <30137.1397057056@sss.pgh.pa.us>
 <CAM-w4HNawF2sPDSUrZsng7zgGnste-Vy1+42AufyVq9UkVeCZw@mail.gmail.com>
 <20140422223230.GL10046@momjian.us>
 <CAM3SWZRbVyijRX0vwhfap1S1o0OM1eGczJsQuL4CCrAs1++z7A@mail.gmail.com>
 <535777DA.3000203@2ndquadrant.com>
 <20140423105630.GQ10046@momjian.us>
Date: Wed, 23 Apr 2014 21:17:49 +0900
Message-ID: 
 <CAB7nPqRGyBfCK9GKr-cqna02Mr5JnqaodQD=q54xttcZn82zRg@mail.gmail.com>
Subject: Re: Re: default opclass for jsonb (was Re: Call for
 GIST/GIN/SP-GIST opclass documentation)
From: Michael Paquier <michael.paquier@gmail.com>
To: Bruce Momjian <bruce@momjian.us>
Cc: Petr Jelinek <petr@2ndquadrant.com>, Peter Geoghegan <pg@heroku.com>,
 Greg Stark <stark@mit.edu>,
 Tom Lane <tgl@sss.pgh.pa.us>, Robert Haas <robertmhaas@gmail.com>,
 Heikki Linnakangas <hlinnakangas@vmware.com>,
 Andrew Dunstan <andrew@dunslane.net>,
 "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Content-Type: multipart/alternative; boundary=001a113439208e980704f7b4b882
Precedence: bulk
Sender: pgsql-hackers-owner@postgresql.org

--001a113439208e980704f7b4b882
Content-Type: text/plain; charset=ISO-8859-1

On Wed, Apr 23, 2014 at 7:56 PM, Bruce Momjian <bruce@momjian.us> wrote:

> On Wed, Apr 23, 2014 at 10:20:42AM +0200, Petr Jelinek wrote:
> > I vote for changing it, even though neither option is ideal I think
> > that given the nature of datatype the current default will break
> > inserts for common usage pattern and that's much worse than not
> > being able to use the index for some operators.
>
> I agree.  We should choose the most general option as the default.
>
+1. Less operators are supported by the now-named jsonb_hash_ops but at
least users won't be surprised by failures caused by too long index records.
-- 
Michael

--001a113439208e980704f7b4b882
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><br><div class=3D"gmail_extra"><br><br><div class=3D"gmail=
_quote">On Wed, Apr 23, 2014 at 7:56 PM, Bruce Momjian <span dir=3D"ltr">&l=
t;<a href=3D"mailto:bruce@momjian.us" target=3D"_blank">bruce@momjian.us</a=
>&gt;</span> wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-=
left:1px solid rgb(204,204,204);padding-left:1ex"><div class=3D"">On Wed, A=
pr 23, 2014 at 10:20:42AM +0200, Petr Jelinek wrote:<br>
&gt; I vote for changing it, even though neither option is ideal I think<br=
>
&gt; that given the nature of datatype the current default will break<br>
&gt; inserts for common usage pattern and that&#39;s much worse than not<br=
>
&gt; being able to use the index for some operators.<br>
<br>
</div>I agree. =A0We should choose the most general option as the default.<=
br></blockquote><div>+1. Less operators are supported by the now-named json=
b_hash_ops but at least users won&#39;t be surprised by failures caused by =
too long index records.<br>
</div></div>-- <br>Michael<br>
</div></div>

--001a113439208e980704f7b4b882--