Thanks Aleksey,

I have submitted a recommendation to amend the documentation for "SET TABLEVALIDATE TO".

There is no need to interrogate the return from TABLEUPDATE() when there is NO BUFFERING in effect - the error message is emitted on the INSERT INTO... command.

It seemed reasonable to ASSuME (always learn differently) that the same would be the case when buffering is in effect. That it is necessary to check the return from TABLEUPDATE() needs, in my opinion, to be mentioned in the Help.

Thanks for your help in getting me to understand what to expect.

I am also taking the position that the failure to error when the record count is only 1 off actual (using SET value of 2) is simply a product of the implementation. In this case, checking AFTER the new record is written yields a correct (uncorrupted) header/size so no error is detectable.
There may be arguments against this implementation, but I will leaves those to others.