Re: Demo: Modelling Cost of Travel Paths Between Towns
Date: 15 Nov 2004 12:19:40 -0800
Message-ID: <4b45d3ad.0411151219.e97f3a4_at_posting.google.com>
> 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?
Reading thru the earlier posts, it seems I have lost context of the thread.
> Ed: Matching that model in a Relational data model is easy enough.
> Neo: This problem didn't not have highly variable data and relationships
> or require a higher degree of normalization. Try adding a different
> property to each city or name two cities the same. RM solution will
> require some changes to schema and code already written against.
> XDb2 solution will be minimally affected.
> Alan: No big deal.
> Neo: Please show how the current RM solution can work if two towns
> are named the same (ie change name of town c to a).
> Alan: The table would be thus: town PK, state PK, main_postal_code PK
What I failed to catch here was:
The above problems resulted because Celko's original one-table schema
represented cities redundantly. It should have had a separate table
1) Alan was not able to rename town c to a in the original RM solution
presented by Celko.
2) The original RM solution requires changes to schema to accomodate
two towns with the same name.