[MySQL] FAKE DUPLICATE ROW

I have this table on my MYSQL 4.0.16 : CREATE TABLE pls_in_row ( row_id int(11) NOT NULL auto_increment, pls_id int(11) default NULL, code int(20) default NULL, description varchar(255) default NULL, valid_from date default NULL, PRIMARY KEY (row_id), UNIQUE KEY pls_row (code,pls_id,valid_from) ) TYPE=MyISAM when I push data into the table with a "INSERT IGNORE" statement, everything works fine as long as 'code' is not longer than 9 digits. When I try to insert records with 'code' longer than 9 digits, mysql does not distinguish anymore in the index... so for example if I perform the following, the second statement does not insert anything as the row was duplicate, but actually is not: INSERT IGNORE INTO pls_in_row SET pls_id = 17, code = 372622701, valid_from = '2004-01-01'; > affeted rows: 1; INSERT IGNORE INTO pls_in_row SET pls_id = 17, code = 3726227013, valid_from = '2004-01-01'; > affected rows: 0; leaving out the 'IGNORE' I get this... > ERROR 1062: Duplicate entry '2147483647-17-2004-02-01' for key 2 so that seems that mysql has created a completely wrong key for this row! my question is: is there a limit for the integer part of the index, or a bug any suggestion would be very much appreciated. cheers alessandro