Hello,

We gave a Postgres 8.4 database running under Linux Centos 5.4 and
defined so :

ENCODING = 'SQL_ASCII'
TABLESPACE = pg_default
LC_COLLATE = 'fr_FR.UTF-8'
LC_CTYPE = 'fr_FR.UTF-8'
CONNECTION LIMIT = -1;

Accessing under Linux is no problem but when we merge addresses under
Windows using f.e. Winword an address like 'Résidence Les Cèdres, bt. H"
becomes : " R‚sidence Les CŠdres, bƒt H " and I can't see how to
interface correctly the two systems.

If you have an idea, I would be grateful.

Regards,

Alain

In <4f17eae9$0$5045$ba620e4c (AT) news (DOT) skynet.be> Alain Reymond wrote:

The postgresql manual advices against using SQL_ASCII as encoding when
storing data that isn't limited to ASCII, and especially so when using a
LC_CTYPE not limited to ASCII.

http://www.postgresql.org/docs/8.4/s...multibyte.html

You probably want to look at adjusting the client encoding in whatever
client software or database driver your using on the client system. Word
probably make implicit assumptions about what character encoding is used for
the indata.

With psql you can check and set the client encoding with the command
\encoding or from the environment variable PGCLIENTENCODING. Example...

$ PGCLIENTENCODING=UTF-8 psql foodb -c "select * from foo;" -o foo-utf
$ file foo-utf
foo-utf: UTF-8 Unicode text

$ PGCLIENTENCODING="ISO 8859-1" psql foodb -c "select * from foo;" -o foo-iso
$ file foo-iso
foo-iso: ISO-8859 text

--
Fredrik Jonson