Re: Modeling Address using Relational Theory


AK wrote:
> it will be a little bit country specific. You could limit your
> application for only one country. Or if you cannot figure out all the
> details for all the countries up front, you might be better off with
> Line1, Line2, Line3. Alternatively you could plan to have country
> specific versions

I understand the Line1, Line2, Line3 implementation, but any model with
those attributes screams "this model is not relational", right? I am
thinking about categories other than performance for times when a good
data modeler would decide not to implement the relational model of the
data (thinking about OLTP [online transaction processing] -- ignoring
BI [business intelligence], OLAP [online analytical processing], and
data warehousing reasons for now).

I'm wondering what a good relational model would be in order to get
more insight into this class of changes between relational model and
implementation. Is there a known classification for the pattern that
is illustrated with "address" in disregarding the relational model when
doing the implementation?

Thanks. --dawn

.