From: =?UTF-8?B?6I2S5LqV5YWD5oiQ?= <n2029@ndensan.co.jp>
To: "'Tom Lane'" <tgl@sss.pgh.pa.us>,
	"'Holger Jakobs'" <holger@jakobs.com>
Cc: <pgsql-admin@lists.postgresql.org>
References: <00e201d8431c$cb38eb40$61aac1c0$@ndensan.co.jp>
 <CAKFQuwYCTmp5t5_7n+qRoKZArnBvv309CSi4XVdONXxQL6OX0Q@mail.gmail.com>
 <00f601d84326$5367cf00$fa376d00$@ndensan.co.jp>
 <006501d84350$37bae6f0$a730b4d0$@msym.fr>
 <012c01d84353$1f8a8d30$5e9fa790$@ndensan.co.jp>
 <15e893e2-1237-27f4-58ba-5ea59499fa3d@jakobs.com>
 <1107956.1648549556@sss.pgh.pa.us>
In-Reply-To: <1107956.1648549556@sss.pgh.pa.us>
Subject: RE: About Unicode IVS
Date: Tue, 29 Mar 2022 20:03:45 +0900
Message-ID: <013501d8435c$a8f1c9e0$fad55da0$@ndensan.co.jp>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
thread-index: 
 AQHjbJJi0v014goMGofTVeHbofwGGgJNxTOuAs0qQHoDP/CCzQHZXpaPAIhlApYBPZ/8AqxgDdaw
Content-Language: ja
Archived-At: 
 <https://www.postgresql.org/message-id/013501d8435c%24a8f1c9e0%24fad55da0%24%40ndensan.co.jp>
Precedence: bulk

thank you for your reply.

In SQL Server, the variant character selector is treated as one =
character with two characters. The collation order is =
Japanese_XJIS_140_CS_AS_KS_WS_VSS_UTF8.

Moto.

-----Original Message-----
From: Tom Lane <tgl@sss.pgh.pa.us>=20
Sent: Tuesday, March 29, 2022 7:26 PM
To: Holger Jakobs <holger@jakobs.com>
Cc: pgsql-admin@lists.postgresql.org; n2029@ndensan.co.jp
Subject: Re: About Unicode IVS

Holger Jakobs <holger@jakobs.com> writes:
> It's totally correct that the two characters are still two characters.
> You would have to normalize the string first, so that the combination=20
> becomes one character.

Yeah.  In principle the normalize() function ought to do this for you.  =
But it doesn't seem to shorten the given example for me; I'm not sure if =
that means the example is incorrect, or if it's a bug in normalize().

u8=3D# select octet_length(U&'\+008FBA' || U&'\+0E0102');  octet_length
--------------
            7
(1 row)

u8=3D# select octet_length(normalize(U&'\+008FBA' || U&'\+0E0102'));  =
octet_length
--------------
            7
(1 row)

			regards, tom lane