The following bug has been logged online:

Bug reference: 1268
Logged by: Kent Tong

Email address: kent (AT) cpttm (DOT) org.mo

PostgreSQL version: 7.4.5

Operating system: RedHat 9

Description: Two different Unicode chars are treated as equal in a
query

Details:

Steps:
1. Create a test database: "createdb -E Unicode -U postgres testdb".
2. Create a test table: "create table testtable (id varchar(100) primary
key);".
3. With JDBC, insert a record whose id contains unicode: "insert into
testtable values(<a unicode char whose code is 0x4e8c>);".
4. With JDBC, try to retrieve a record whose id contains a different unicde:
"select from testtable where id=<a unicode char whose code is 0x4e94>;". It
should not find any record but it finds the record created in step 3.

Here is the JUnit test case:

public class PgSQLTest extends TestCase {
private Connection conn;
protected void setUp() throws Exception {
conn = makeConnection();
}
protected void tearDown() throws Exception {
conn.close();
}
public void testChinese() throws Exception {
deleteAll();
insertRow();
PreparedStatement st =
conn.prepareStatement("select * from testtable where id=?");
try {
st.setString(1, "\u4e94");
ResultSet rs = st.executeQuery();
assertFalse(rs.next());
} finally {
st.close();
}
}

private void insertRow() throws SQLException {
PreparedStatement st =
conn.prepareStatement("insert into testtable values(?)");
st.setString(1, "\u4e8c");
st.executeUpdate();
st.close();
}
private void deleteAll() throws SQLException {
PreparedStatement st = conn.prepareStatement("delete from testtable");
st.executeUpdate();
st.close();
}
private Connection makeConnection()
throws ClassNotFoundException, SQLException {
Class.forName("org.postgresql.Driver");
Properties properties = new Properties();
properties.put("user", "postgres");
properties.put("password", "");
return DriverManager.getConnection(
"jdbcostgresql://localhost/testdb",
properties);
}
}



---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo (AT) postgresql (DOT) org)

"PostgreSQL Bugs List" <pgsql-bugs (AT) postgresql (DOT) org> writes:
This would be a matter to take up with the maintainer of your locale
(which you didn't mention, but in any case it's a locale bug). We
just do what strcoll() tells us.

Note that it's possible this is a configuration error and not an
outright bug. Check to make sure that the locale you initdb'd
under is actually designed to work with UTF-8 data.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match