Just my personal bias--I'd rather track the record pointer using the primary key instead of RECNO(). Good point about the recno, though.

>If all updates are buffered, than before tableupdate() your recno() will show negative numbers (buffered records have recno from -1,-2 .. and so on).
>And after tableupdate all those negative records are gone, so you should save your recno BEFORE changing the records...