>>How are developers handling the problem where a primary key and data record is added to a table. The user decides to start the record over so they delete the record. When they start to add the record again with the same key, they will get a duplicate key message because the key, although deleted, is still in the index, until I pack the file. One idea would be to change the key when I delete the record. Is there a better approach anyone is using.
>
>Don't recirculate primary keys - once you've assigned a primary key, it's gone and no longer usable, even if you never actually post it through. In most cases, I'd use a surrogate key - an artificial key used only for relation handling - and discard it even if it nver posted to the table.

That is how we handle it. NewKey(), a stored procedure, supplies the default value, looks up the last key in KeyList.dbf, verifies it and increments, never recirculates.