>True - and I'll try not to start any discussions on whether or not that is correct behaviour
I know what you're saying and couldn't agree more.

> - the intent is to avoid moving the file pointer so that deletes and changes don't accidently get committed when they shouldn't. I believe that is what happens >when the pointer is moved even in a buffered file. But I sit to be corrected (the shock ya know *g*).

This is why I always use Optimistic Table Buffering and TABLEUPDATE(0) for single record updates and TABLEUPDATE(1) for multi. This way record pointer movement won't accidentaly update the table. Have a look at Eriks last suggestion. Looks promising. Thanks for all the suggestions.