>>Thanks Hilmar,
>>True, our current message should say something else - but saying what you propose won't be true either, because the record IS saved, and that was what my main question was about.
>>OTOH, do you have any scenario when the tableupdate(), in the specified conditions, might return false?
>
>You gave the scenario yourself: a corrupted index. Really, I don't know whether this is related or not, but just consider the possibility that the corrupted index is the cause, rather than the result of your problems.

That is actually what I meant, that the index corruption is the cause - but I am not 100% sure if is the case...