> Thanks Kris and Oliver for your explanations.
> Why I want to set the client_encoding to latin1? That's a fair
> question. Unfortunately I removed the statement following the COPY
> command to keep the posting short:
> SET CLIENT_ENCODING TO 'LATIN1';
> COPY temp_test FROM '/home/postgres/temp_test.dat';
> I want to load a latin1-encoded file using the COPY command.
Hmm. I am suprised that COPY does not let you specify the encoding of
the input file. Using client_encoding for this seems wrong: there are
some cases it can't handle, e.g. trying to load a LATIN1-encoded file
into a table that has a name that can't be represented in LATIN1.
I suppose that in the absence of backend support for this, we could add
some URL parameter that allows client_encoding to be changed, with
suitably dangerous warnings around using it. Then you can temporarily
flip client_encoding to LATIN1 for the duration of the COPY, and revert
it to UNICODE afterwards.
In response to
pgsql-jdbc by date
|Next:||From: Markus Schaber||Date: 2004-11-04 10:30:56|
|Subject: Re: Avoiding explicit addDataType calls for PostGIS|
|Previous:||From: mbch67||Date: 2004-11-04 07:08:43|
|Subject: Re: 8.0.0beta4: "copy" and "client_encoding"|