[AccessD] Normalization discussion

Susan Harkins ssharkins at gmail.com
Sat Aug 30 15:09:11 CDT 2014


>
>  For example, every animal has a father and mother.  You may not know it
> because it was an acquisition or that it came from law enforcement, but it
> does have one, you just don't know it.  That's where a null comes in and is
> the difference between not having an attribute and not knowing the value
> for
> it.   So me, I would have a sire and dam FK fields in the animal record and
> they remain null for anything other than a live birth.
>

===========An excellent idea and one that I hadn't considered. Thank you.

>
>  All records in a table should have the same "shape".  That is, you should
> be able to fill in a value for every column (assuming you know it).   If
> you
> cannot to that for a record, then your describing more than one "thing" in
> your table.
>

===========This is already done and I'm comfortable with the table
structure.

>
>  I think by the time you make up your list, break things up into groups,
> you'll find it's just common sense.
>
>  But to forgo normalization entirely will give you a database that is
> difficult to work on, have poor performance, and other issues.
>
>
===========I'm not going to do that. :) But, I am feeling better about
taking small shortcuts -- just like the one you mentioned above. I was
making it way too hard.

Thanks!

Susan H.


More information about the AccessD mailing list