Re: [HACKERS] Troubles using German Umlauts with JDBC

Rene,

Two comments on your writeup about the problem:

1) Depending on version you will see different behavior:
7.0 - default client character set is used
7.1 - database character set is used (although it may be reported
incorrectly as SQL_ASCII)
7.2 - database character set is used if multibyte, else use the
client character set.

In all versions it is possible to set the character set explicitly via
the charSet parameter.

2) The charSet parameter (as can any parameter the driver expects) can
also be set in the connection URL. (i.e.
jdbc:postgresql://localhost/dbname?charSet=UTF-8&user=foo&password=bar
shows passing the charSet, user and password in the URL)

thanks,
--Barry

Rene Pijlman wrote:
> I've added a new section "Character encoding" to
> http://lab.applinet.nl/postgresql-jdbc/, based on the
> information from Dave and Barry.
>
> I haven't seen a confirmation from pgsql-hackers or Bruce yet
> that this issue will be added to the Todo list. I'm under the
> impression that the backend developers don't see this as a
> problem.
>
> Regards,
> René Pijlman
>
> On Tue, 04 Sep 2001 10:40:36 -0700, Barry Lind wrote:
>
>>I would like to add one additional comment. In current sources the jdbc
>>driver detects (through a hack) that the server doesn't have multibyte
>>enabled and then ignores the SQL_ASCII return value and defaults to the
>>JVM's character set instead of using SQL_ASCII.
>>
>>The problem boils down to the fact that without multibyte enabled, the
>>server has know way of specifiying which 8bit character set is being
>>used for a particular database. Thus a client like JDBC doesn't know
>>what character set to use when converting to UNICODE. Thus the best we
>>can do in JDBC is use our best guess (JVM character set is probably the
>>best default), and allow the user to explicitly specify something else
>>if necessary.
>>
>>thanks,
>>--Barry
>>
>>Rene Pijlman wrote:
>>
>>>[forwarding to pgsql-hackers and Bruce as Todo list maintainer,
>>>see comment below]
>>>
>>>[insert with JDBC converts Latin-1 umlaut to ?]
>>>On 04 Sep 2001 09:54:27 -0400, Dave Cramer wrote:
>>>
>>>
>>>>You have to set the encoding when you make the connection.
>>>>
>>>>Properties props = new Properties();
>>>>props.put("user",user);
>>>>props.put("password",password);
>>>>props.put("charSet",encoding);
>>>>Connection con = DriverManager.getConnection(url,props);
>>>>where encoding is the proper encoding for your database
>>>>
>>>>
>>>For completeness, I quote the answer Barry Lind gave yesterday.
>>>
>>>"[the driver] asks the server what character set is being used
>>>for the database. Unfortunatly the server only knows about
>>>character sets if multibyte support is compiled in. If the
>>>server is compiled without multibyte, then it always reports to
>>>the client that the character set is SQL_ASCII (where SQL_ASCII
>>>is 7bit ascii). Thus if you don't have multibyte enabled on the
>>>server you can't support 8bit characters through the jdbc
>>>driver, unless you specifically tell the connection what
>>>character set to use (i.e. override the default obtained from
>>>the server)."
>>>
>>>This really is confusing and I think PostgreSQL should be able
>>>to support single byte encoding conversions without enabling
>>>multi-byte.
>>>
>>>To the very least there should be a --enable-encoding-conversion
>>>or something similar, even if it just enables the current
>>>multibyte support.
>>>
>>>Bruce, can this be put on the TODO list one way or the other?
>>>This problem has appeared 4 times in two months or so on the
>>>JDBC list.
>>>
>>>Regards,
>>>René Pijlman <rene(at)lab(dot)applinet(dot)nl>
>>>
>