Re: Restoring wrong accents

=?ISO-8859-1?Q?Jos=E9_Roberto_Motta_Garcia?= <garcia@cptec.inpe.br> writes:
> We're having problems restoring the characters correctly here. We have a
> 8.1 PG on Suse created with LATIN1 and pt_BR. As a latin language we use
> several special chars like �, �, �, �, etc.

> We made a dump -a and after restoring it to an also 8.1 PG on Suse
> created with LATIN1 and pt_BR we noticed that a lot of these special
> chars weren't restored correctly. We couldn't notice any rule for that,

It seems highly unlikely that you actually had the same encoding in both
cases, since if that were the case no conversion would have occurred.
Please double-check the two database encodings.  Also, what encoding is
called out in the first few lines of the dump file?

            regards, tom lane