On Saturday 05 June 2010 7:07:35 am Wang, Mary Y wrote:
> Thanks Alban, but I'm still confused.
> I just read the PostgreSQL manual, and got the impression that UTF8 would
> take care of ALL character sets, so I put SET client_encoding = 'UTF8' in
> the dump file, and reload it again, then I got even more errors like:
> psql:/tmp/060410finalsr15dump.txt:31430: ERROR: invalid byte sequence for
> encoding "UTF8": 0x93 psql:/tmp/060410finalsr15dump.txt:31795: ERROR:
> invalid byte sequence for encoding "UTF8": 0x93
> psql:/tmp/060410finalsr15dump.txt:31826: ERROR: invalid byte sequence for
> encoding "UTF8": 0x85 psql:/tmp/060410finalsr15dump.txt:31827: ERROR:
> invalid byte sequence for encoding "UTF8": 0x92
> psql:/tmp/060410finalsr15dump.txt:31874: ERROR: invalid byte sequence for
> encoding "UTF8": 0x93 psql:/tmp/060410finalsr15dump.txt:32111: ERROR:
> invalid byte sequence for encoding "UTF8": 0x93
> psql:/tmp/060410finalsr15dump.txt:32112: ERROR: invalid byte sequence for
> encoding "UTF8": 0x92
>
> What is the client encoding that I should set to so that it would take care
> all character sets?
> http://www.postgresql.org/docs/8.3/interactive/multibyte.html
>
> Any suggestions?
> Thanks
> Mary
>
I am afraid it does not work that way. You have to think translation tables. In
order for the process to work Postgres needs to know the correct encoding it is
receiving in order to make the translation to UTF8. Much the same problem as
giving someone who knows how to translate French to English, a document that
contains a mixture of words in different languages. If they don't know what the
words are they cannot be translated.
--
Adrian Klaver
adrian.klaver@gmail.com