If the database has been configured to support duplicate records, the method will remove all of the duplicate records. To remove individual duplicate records, you must use a Berkeley DB cursor interface.