java version "1.3.0rc2"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.3.0rc2-Y)
Java HotSpot(TM) Client VM (build 1.3.0rc2-Y, mixed mode)

There is a test program:

--- Test.java

import java.sql.*;
import java.util.*;

public class Test
{
 static void dump(byte[] buf)
 {
  for(int i=0; i<buf.length; i++)
    {
     int v = buf[i] & 0xff;
     System.out.print( Character.forDigit(v >> 4,16) );
     System.out.print( Character.forDigit(v & 0xf,16) );
    }

  System.out.println();
 }

 public static void main(String[] args) throws Exception
 {
  Class.forName("sun.jdbc.odbc.JdbcOdbcDriver");

  Properties connInfo = new Properties();
  connInfo.put("charSet","Cp1251");

  Connection db = DriverManager.getConnection("jdbc:odbc:test", connInfo);

  Statement stmt = db.createStatement();

  stmt.executeUpdate("DROP TABLE Test");
  stmt.executeUpdate("CREATE TABLE Test (Val VARCHAR(10))");

  stmt.executeUpdate("INSERT INTO Test (Val) VALUES (\'\u0410\u0411\')");

  ResultSet rs = stmt.executeQuery("SELECT Val FROM Test");

  while( rs.next() )
    {
     dump( rs.getBytes(1) );
    }

  Statement rsstmt = db.createStatement(ResultSet.TYPE_SCROLL_INSENSITIVE,
				      ResultSet.CONCUR_UPDATABLE);

  rs = rsstmt.executeQuery("SELECT Val FROM Test");

  rs.first();

  rs.updateString(1,"\u0410\u0411");
  rs.updateRow();
  rs.close();
  rsstmt.close();

  rs = stmt.executeQuery("SELECT Val FROM Test");

  while( rs.next() )
    {
     dump( rs.getBytes(1) );
    }
 }
}
---

Create new database with MS Access 97, and execute this.

The program output:

---
c0c1
d090d091
---

Similar results produced, when using another methods instead
updateString(). For example, if replace updateString() with:

  byte[] buf = "\u0410\u0411".getBytes("Cp1251");
  rs.updateBytes(1,buf);

then output is:

---
c0c1
00fd7600582275001100
---

And, if use:

  byte[] buf = "\u0410\u0411".getBytes("Cp1251");
  rs.updateBinaryStream(1,new java.io.ByteArrayInputStream(buf),buf.length);

then output is:

---
c0c1
1cfd7600582275001000
---

Correct behavoir (in all these cases) should be:
---
c0c1
c0c1
---
(Review ID: 103243) 
======================================================================

The only workaround - is using INSERT and UPDATE statements
======================================================================

Sent to Merant.

  xxxxx@xxxxx   2000-05-03

This is not fixed I've hebrew char in DB, then it's 
impossible for me to use TYPE_SCROLL_INSENSITIVE cause the 
char return by getString are wrong,
the workaround is to return JDBC 1.0 !?!?!?!?
if you have any answers thanks

Hello Sun!
I was believed that evething is fixed, until today i run 
some tests.
The bug is NOT fixed.
It's still exist in ladybird (JDK 1.3.1 build 24) and in 
merlin-beta (JDK 1.4.0-beta-b65).

It's not fixed in ladybird so I will create a new bug 
report.

It's seems to already resubmited. See bug 4476228

Finally resubmit - vote for bug 4486195.