This option is normally used where a number of dimension keys will fail to join and we wish to automatically add new dimension records. The first phase of the generated code will perform a set based join and insert of the data. The dimensions that failed to join will be assigned the zero key value. The second phase of the procedure will use a cursor to cycle through all the stage records that have a zero key dimension join. By setting the appropriate flags in the dimension joins when building this procedure we can request that any missing dimension records be automatically added.