Simplifies Chines(GB2312) character issue- Solaris Weblogic 10.3

Question

i am trying to fetch and update simplified chinese character(GB2312) from database , Update part is works fine in weblogic 10.3 windows machine , but it fails(junk character) in weblogic 10.3 Solaris machine , but Fetching and displaying chines character is working fine in both environments

Fetching DAO

  while (rs.next()) {           
                    Base64 base64 = new Base64();
                    byte[] notesByte = base64.encode(rs
                            .getBytes("notes"));                
    }

UI (Android)

     byte[] notes= Base64.decode(notesByteStr , Base64.DEFAULT);
       notesText.setText(new String(notes, "GB2312")); // Displaying chines char

     notesByte=  Base64.encode(notesText.getText().toString().trim().getBytes("GB2312"),  Base64.DEFAULT) // send to db 

Update DAO

 getSession().createSQLQuery(notesUpdateQuery)
                    .setParameter(0, new String(base64.decode(notesByte)))
                    .executeUpdate();

note: Source txt file encoding : UTF-8

Answer 1

Well yes, look at this in your Update DAO:

new String(base64.decode(notesByte))

That's using the platform default encoding to convert the base64-decoded byte array into a string. The base64-decoded byte array is actually text encoded in GB2312 - not in the platform default encoding.

To convert it properly, you need to specify the same encoding everywhere:

new String(base64.decode(notesByte), "GB2312")