Re: looking for old Datamation poster on Normalizing

"-CELKO-" <jcelko212_at_earthlink.net> wrote in message news:1156608184.171579.223640_at_p79g2000cwp.googlegroups.com...
> >> I've allways been disturbed by the presentation of 4NF in the "puppies
& tricks" poster. It seems to me that there's a misprint, somewhere, or that the key of one of the tables has undergone a shift between the 3NF example and the 4NF example. <<
>
> Marc Rettig is the author. He takes the 3NF example and adds a
> "costume" attribute to the "Puppy Tricks" table, but mulitple costumes
> can be assigned to each puppy, so you split out a "Puppy Costumes"
> table which is all key (puppy_nbr, costume).
>

I think that's where the problem lies...IIRC. It seems to me that if multiple costumes can be assigned to each puppy, then, after adding the costume attribute to the puppy tricks table, that table is not only not in 4NF, but it isn't in 1NF either. Received on Sat Aug 26 2006 - 19:31:57 CEST