Re: Easy Modeling Qs

From: JRStern <JXSternChangeX2R_at_gte.net>
Date: Tue, 02 Jul 2002 01:53:02 GMT
Message-ID: <3d2107c6.43043783_at_news.verizon.net>


On Tue, 2 Jul 2002 01:24:31 +0000 (UTC), "Alex Petrov" <master.db_at_mail.ru> wrote:
>>1)In 2 tables with a 1:1 relationship, I think I could combine them
>>into 1 table but think it would be more efficient to have 2.
>
>Yes, this technique is called "vertical partitioning" ((opposite to
>"horizontal partitioning" based on age/use criteria when 2 or more
>tables have absolutely the same columns but store different rows (e.g
>"Sales_2000" and "Sales_2001" tables that store rows for sales made by
>company in 2000 and 2001 accordingly))

Well, he didn't really say whether he was doing vertical or horizontal, but it's (probably!) one or the other.

>>I have a table that is a 1:n to 3 different tables...can I do that? I am
>>used to association tables for many to many relationships but these
>>are all 1:n. Seems like maybe this is bad design but I can't figure
>>out a better way.
>
>That's ok, associative entities/tables always used this way --to
>eliminate many-to-many relationships by resolving/changing them into 2
>or more one-to-many relationships.

Yes, in effect the parent table *is* the association table, you just have to change your perception (or break out the fields into a separate (horizontal!?) table).

J. Received on Tue Jul 02 2002 - 03:53:02 CEST

Original text of this message