Re: We claim that delete anomality is due to table not being in 3NF, but...

From: Jon Heggland <>
Date: Thu, 30 Oct 2008 08:53:56 +0100
Message-ID: <gebp6k$dhh$> wrote:
> So in essence, we don't consider just the loss of existence an
> anomaly?

I wouldn't (and didn't) put it that way without defining precisely what I mean by "existence". Are you talking about the existence of a teacher as such? Or about the existence in the table of some particular name? if so, why single out the TEACHER_NAME? Why not consider the "loss of existence" of the SUBJECT_ID or SUB_NAME?

The loss of the last appearance of some TEACHER_NAME value does not imply that the corresponding teacher does not exist. The table does not talk about the existence of teachers, only what subjects are taught by what teachers. You can imply the existence of a teacher from the table, but not the non-existence, despite the closed-world assumption. Thus, if a teacher does not teach any subjects, it is perfectly correct (i.e. not anomalous) that hir name does not appear in the table.

Received on Thu Oct 30 2008 - 08:53:56 CET

Original text of this message