Codepage conversion from multi-lingual table using JDBC

Question

Hello,
I have an Oracle database that has been around since before any unicode was available. It is currently 9.2.0.7.0 and the charset is WE8ISO8859P1. I am trying to port all of the data in the 'Translation' table to a UTF-8 database. The problem is that each of the translated strings is stored in its own native MS Windows codepage. This table was populated over the years via OCI connections in C/C++, and this method worked. But now we need it moved to UTF-8 and are using JAVA. Obviously, all the cp1252 languages come over just fine, but all the other languages get converted to garbage by JDBC, since it is auto-converting them and assumes the source is 8859. I could do it with OCI, as I mentioned, but I am mandated to use JAVA and a web interface, so users can choose what data to move/convert to the new DB (on-demand) from a web page.
Is there a way to force the Oracle connection to think the database is in a different codepage so that the conversion comes out correct? OR
Is there a way to get the data from Oracle as unconverted chars, or generic binary, or something, and then have JAVA convert it to UTF-8 from the correct codepage before writing it to the new table?
Any help is appreciated.
Monte
Message was edited by:
Wookie

Answer 1

## Is there a way to force the Oracle connection to think the database is
## in a different codepage so that the conversion comes out correct?
No.
## Is there a way to get the data from Oracle as unconverted chars,
## or generic binary, or something, and then have JAVA convert it
## to UTF-8 from the correct codepage before writing it to the new table?
Yes. Use:
SELECT UTL_RAW.CAST_TO_RAW(col) FROM tab;
and retrieve the value with ResultSet.getBytes. Then convert the retrieved byte[] to java.lang.String using String(byte[] bytes, String charsetName) constructor. Note, charsetName is Java name. Then, you can use setString() on PreparedStatement containing the INSERT for the converted value.
-- Sergiusz