Re: Demo: Modelling Cost of Travel Paths Between Towns

On 14 Nov 2004 17:52:39 -0800, Neo wrote:

>> > Yes, a separate table for states is the correct way to avoid redundant
>> > PA's (even if PA doesn't have any properties). Alan didn't provide
>> > such a schema thus he has redundant PA's.
>>
>> States do not need a seperate table because (at this point, anyway)
>> they have no properties.
>
>A separate table for states is need to avoid redundant PA's. Your
>current schema has redundant PAs. The first PA represent PA. The
>second PA also represents the same PA as the first. According to CJ
>Date, redundancy leads to update anomalies. Assuming no
>auto-sychronization, your db can be corrupted by modifying one of the
>PAs. Please reread chapter 10 and in particular pg 312 of "An Intro to
>Db Systems" 6th Ed.

Hi Neo,

So if state boundaries are changed and the city of Unreal, PA will henceforth be part of Texas, changing the row for Unreal, PA to Unreal, TX would corrupt the db?

Hmmmm, interesting...

Best, Hugo

-- 

(Remove _NO_ and _SPAM_ to get my e-mail address)