Re: Proposal: 6NF
From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Sat, 07 Oct 2006 12:24:46 GMT
Message-ID: <i0NVg.1604$cz.23912_at_ursa-nb00s0.nbnet.nb.ca>
>
>
> That will work, and I already tried to say as much. You can have two
> views, one with all the animals and only the columns that are common to
> all the animals (the inner union suggested by Marshall), and one with
> the single attribute you need for this query plus a key that is shared
> by all animals. As a general principle, you'll end up with an inner
> union view for each class containing it and all of its subtypes, and
> you'll also need a similar view for each attribute containing all of the
> types it appears in. If you apply multiple inheritance willy nilly,
> eggs_laid could be used with robotic chicken, which are machines, not
> animals
Date: Sat, 07 Oct 2006 12:24:46 GMT
Message-ID: <i0NVg.1604$cz.23912_at_ursa-nb00s0.nbnet.nb.ca>
Sampo Syreeni wrote:
>> SELECT COUNT(1) >> FROM Animals >> ANTIJOIN Layers >> ON Animals.id = Layers.id >> WHERE Layers.eggs > 5
>
>
> That will work, and I already tried to say as much. You can have two
> views, one with all the animals and only the columns that are common to
> all the animals (the inner union suggested by Marshall), and one with
> the single attribute you need for this query plus a key that is shared
> by all animals. As a general principle, you'll end up with an inner
> union view for each class containing it and all of its subtypes, and
> you'll also need a similar view for each attribute containing all of the
> types it appears in. If you apply multiple inheritance willy nilly,
> eggs_laid could be used with robotic chicken, which are machines, not
> animals
Um, so if you suggest doing something incredibly stupid, you will devise a straw man that is incredibly unpersuasive.
[remaining nonsense snipped] Received on Sat Oct 07 2006 - 14:24:46 CEST