JDBC program is inserting ¿ in NVARCHAR2 column.

Question

I have a 10.2.0.1 database on a windows XP machine.
The value of NLS_CHARACTERSET is WE8MSWIN1252 and the value of NLS_NCHAR_CHARACTERSET     is AL16UTF16.
I have a table with a NVARCHAR2 column. I am using JDBC thin driver to connect to the database. When I am to inserting a japanese character to this column, the value that is getting inserted is ¿ (inverted question mark) instead of the actual Japanese character. Following is the code snippet of my Java program
Class.forName("oracle.jdbc.OracleDriver");
props.put("user", "bcan");
props.put("password","bcan2226");
url = "jdbc:oracle:thin:@myserver:1521:eng";
Connection conn = DriverManager.getConnection(url, props);
String value = "abcdefghi\u7c73"; //\u7c73 is the japanese character
String sql = "insert into I18N VALUES ('" + value + "')";
Statement stmt = conn.createStatement();
stmt.executeUpdate(sql);
Can you please let me know why is ¿ getting inserted instead of the japanese character I am trying to insert?
BTW I noticed that if I use PreparedStatement like the following then the proper character gets inserted.
Class.forName("oracle.jdbc.OracleDriver");
props.put("user", "bcan");
props.put("password","bcan2226");
props.put("oracle.jdbc.defaultNChar","true");
url = "jdbc:oracle:thin:@myserver:1521:eng";
Connection conn = DriverManager.getConnection(url, props);
String value ="abcdefghi\u7c73";
String sql = "insert into I18N VALUES (?)";
PreparedStatement stmt = conn.prepareStatement(sql);
stmt.setString(1,value);
stmt.executeUpdate();
Thanks
Sudipta

Answer 1

I do not think that you are hitting a "bug". Instead what you are seeing is expected behaviour since sql and literals (e.g. your string) are processed in the database character set unless you specify otherwise (e.g. explicit bind).
Go through docs per links in my previous post.
For the bug referenced, it seemed to be fixed according to jdbc driver Readme of 10.2.0.4.