Re: Relational Design

In article <HCFb5.16400$7%3.1011193_at_news.flash.net>, nospam_at_nospam.com says...
> I like the idea of enforcing a unique address at the time data entry,
>
> Would it be better to divide up the address components into columns and use
> the combination as a unique key? Or is there a performance hit on the
> database with this method?
>

Conventional normalization rules would put addresses and volunteers in separate tables, with 1:N between addresses and volunteers. Either way, you'd need to test for an extant (duplicate) address at data entry time to avoid breaking referential integrity between the two entities.

As to splitting address components out into columns, this will produce a performance hit, but its probably negligible unless the rowcount is _really_ large. Received on Fri Jul 14 2000 - 00:00:00 CEST