>>I know you dont want a lecture, but you are aware that table buffering does not prevent you from treating it - programmatically - as if it was row-buffered I hope.
>
>Jim,
>
>Interesting... I may not be aware. Do you mean using the nRows parameter in the TableUpdate while the buffering is set to table?
>
>Renoir

Maybe using nRow... the main thing is to program to allow only 1 record at a time to be in 'edit mode' and/or to actually TableUpdate each new/revised child record as it is clean and ready.
Doing it this way lets the pointer move without consequence because you and only you control the actual TableUpdate (all conditions causing an implicit TableUpdate do not apply to table buffering).

Good luck
JimN