? (question mark) characters

Ok, I tried all the drivers currently available at http://jdbc.fastcrypt.com.
They all seem to correct the problem with long (>8k) SQL statements, but
they're all broken considering Portuguese characters like áéíóú (aeiou with
accent)... they are replaced by question marks (?)... any ideas? Should I try
to compile my own drivers? How?
BTW I am using JDK 1.3 from Sun (Solaris and Linux)...

--
Ricardo Pardini
rpardini(at)organox(dot)com(dot)br

> Ok, I tried all the drivers currently available at http://jdbc.fastcrypt.com.
> They all seem to correct the problem with long (>8k) SQL statements, but
> they're all broken considering Portuguese characters like áéíóú (aeiou with
> accent)... they are replaced by question marks (?)... any ideas? Should I try
> to compile my own drivers? How?
> BTW I am using JDK 1.3 from Sun (Solaris and Linux)...

I suspect the problem is not with the jdbc driver but with the way you
have configured postgres, try building postgres with:

./configure --with-java --enable-multibyte=UNICODE --enable-unicode-conversion

And use initdb with the -E UNICODE option:

initdb -E UNICODE -D [path to data directory]

And see if that works, good luck!

Michael

Ricardo Pardini wrote:
> they're all broken considering Portuguese characters
> like áéíóú (aeiou with accent)... they are replaced
> by question marks (?)... any ideas?

What's the character encoding of the database? See
http://postgresql.demunnikservices.nl/users-lounge/docs/7.1/admi
n/multibyte.html (this documentation is about character
encoding, not just multibyte support like the title suggests).
And see also
http://fts.postgresql.org/db/mw/msg.html?mid=1029464 about how
multibyte support relates to character encoding support.

Are these special characters converted when you store them in
the database, or when you retrieve them from he database?

>Should I try to compile my own drivers?

No, that won't fix it.

Regards,
René Pijlman

Thank you all for your help:

Dia Thursday, 30 de August, 2001 10:42, Rene Pijlman wrote:
> > like áéíóú (aeiou with accent)... they are replaced
> > by question marks (?)... any ideas?
> What's the character encoding of the database?

It doesn't matter. I have tried it with SQL_ASCII, ISO-8859-1, and with an
installation on Solaris which doesn't have encodings ("psql -l" doesnt list
the Encoding field).

Strange thing is, if I use the jdbc7.0-1.2.jar from
http://jdbc.postgresql.org, the special characters work ok. But then I have
the 8k statement limitation problem.

> Are these special characters converted when you store them in
> the database, or when you retrieve them from he database?

They're converted when I insert. I can insert the same characters directly
into psql (into the same table) and it works ok (I can retrieve them without
problems). I can use LIBPQ apps (PHP, PGACCESS etc) and it works ok with all
backends (running with encodings SQL_ASCII, ISO, etc)...

--
Ricardo Pardini
rpardini(at)organox(dot)com(dot)br

Ricardo,

Is your database compiled for multibyte support? And what character set
is the database you are connecting to created with? (to get the
database character set do a 'psql -l'). If the answer to the first
question is no or the answer to the second question is 'SQL_ASCII', then
only 7bit characters are supported.

thanks,
--Barry

Ricardo Pardini wrote:
> Ok, I tried all the drivers currently available at http://jdbc.fastcrypt.com.
> They all seem to correct the problem with long (>8k) SQL statements, but
> they're all broken considering Portuguese characters like áéíóú (aeiou with
> accent)... they are replaced by question marks (?)... any ideas? Should I try
> to compile my own drivers? How?
> BTW I am using JDK 1.3 from Sun (Solaris and Linux)...
>
>

On Thu, 2001-08-30 at 18:41, Barry Lind wrote:
> Ricardo,
>
> Is your database compiled for multibyte support? And what character set
> is the database you are connecting to created with? (to get the
> database character set do a 'psql -l'). If the answer to the first
> question is no or the answer to the second question is 'SQL_ASCII', then
> only 7bit characters are supported.
:
> > Ok, I tried all the drivers currently available at http://jdbc.fastcrypt.com.
> > They all seem to correct the problem with long (>8k) SQL statements, but
> > they're all broken considering Portuguese characters like áéíóú (aeiou with
> > accent)... they are replaced by question marks (?)... any ideas? Should I try
> > to compile my own drivers? How?
> > BTW I am using JDK 1.3 from Sun (Solaris and Linux)...

For portugese you should be using -E LATIN1

Cheers

Tony Grant

--
RedHat Linux on Sony Vaio C1XD/S
http://www.animaproductions.com/linux2.html
Macromedia UltraDev with PostgreSQL
http://www.animaproductions.com/ultra.html

On Thu, 30 Aug 2001 11:06:13 -0300, you wrote:
>Thank you all for your help:
>Dia Thursday, 30 de August, 2001 10:42, Rene Pijlman wrote:
>> > like áéíóú (aeiou with accent)... they are replaced
>> > by question marks (?)... any ideas?
>> What's the character encoding of the database?
>
>It doesn't matter. I have tried it with SQL_ASCII, ISO-8859-1, and with an
>installation on Solaris which doesn't have encodings ("psql -l" doesnt list
>the Encoding field).

Have you tried it with a database which was created with -E
LATIN1 and with an installation configured with
--enable-multibyte?

If so, can you post a small Java program that reproduces the
problem?

Regards,
René Pijlman <rene(at)lab(dot)applinet(dot)nl>

* Rene Pijlman <rene(at)lab(dot)applinet(dot)nl> wrote:
|
| Have you tried it with a database which was created with -E
| LATIN1 and with an installation configured with
| --enable-multibyte?

You don't need multibyte for iso-8859-1. A simple test case reproducing the
problem would be nice though. I don't have any 7.1 servers in production,
but the 7.0 series doesn't have any problems at all with iso-8859-1.

--
Gunnar Rønning - gunnar(at)polygnosis(dot)com
Senior Consultant, Polygnosis AS, http://www.polygnosis.com/

On 02 Sep 2001 03:35:29 +0200, you wrote:
>You don't need multibyte for iso-8859-1.

That's what I thought. But with current CVS (7.2) creating a
database with -E LATIN1 fails without multibyte support. See the
link in one of my previous postings in this thread.

Regards,
René Pijlman <rene(at)lab(dot)applinet(dot)nl>