Обсуждение: iso-8859-1 and utf-8

Hi.

I found out by accident that my db-dump, at least in ver. 8.3.3,
automagically converts from iso-8859-1 to utf-8. Our db has been
around since January 2000 and iso-8859-1 was chosen back then as
encoding. When I occasionally imported the nightly dump to a test-db
it would complain very early during the import and abort if the
encoding was incorrect. I think this was also the case with 8.3.1
although I'm not certain.

Yesterday I did another import on my test-db and this was created with
utf-8 as encoding. The import did not complain and I ended up with an
utf-8-db which is what we have aimed at in the near future. Our web
had some iso-8859-1-quirks which we have solved.

I've started to plan a migration to utf-8. I assume it requires a
complete db-dump, delete the iso-8859-1-db, create an utf-8-db and do
an import. Previously I had piped the dump through iconv to convert
but this does not seem to be a requirement any longer.

--
regards
Claus

When lenity and cruelty play for a kingdom,
the gentler gamester is the soonest winner.

Shakespeare

On Thu, Sep 25, 2008 at 1:59 AM, Claus Guttesen <kometen@gmail.com> wrote:
> Hi.
>
> I found out by accident that my db-dump, at least in ver. 8.3.3,
> automagically converts from iso-8859-1 to utf-8. Our db has been
> around since January 2000 and iso-8859-1 was chosen back then as
> encoding. When I occasionally imported the nightly dump to a test-db
> it would complain very early during the import and abort if the
> encoding was incorrect. I think this was also the case with 8.3.1
> although I'm not certain.

I'm guessing that the client encoding is set to utf-8 instead of
iso-8859-1 on the client end.  pg automatically converts to the client
encoding of the user connecting / environment.

I know that at least 8.2 supports the -E switch to set client encoding.

psql --help
MUCH SNIPPED:
-E, --encoding=ENCODING     dump the data in encoding ENCODING