Re: erd to db

Thank you.

Yves

--
Yves Guerin
CUSM-MUHC
Groupe Interfaces
Montreal - Canada
"Jan Hidders" <hidders_at_uia.ua.ac.be> wrote in message
news:3ca0a933$1_at_news.uia.ac.be...


> A small addition and some small corrections.


>

> "Jan Hidders" <hidders_at_uia.ua.ac.be> wrote in message

> news:3ca08d65$1_at_news.uia.ac.be...

> >

> > - Every multivalued attribute of entity or relationship gets its own

> table:

> >     1. The columns consist of the primary key the table for the

> > entity/relationship plus a column with the name of the attribute.

> >     2. The primary key of this table consists of all its columns. Note

> that

> > you may end up with tables and/or columns with the same name, in which

> case

> > you have to disambiguate them by for example prefixing them with the

> entity

> > type name or relationship name.

>

>     3. The columns of point 1. are indicated as a foreign key to the table

> for the entity/relationship, if it exists. If it does not exist then the

> target becomes the table where the relationship is represented by a

foreign


> key. (This happens for certain identifying relationships and


one-to-many/one


> relationships.)

>

> > - Non-identifying one-to-many relationships do *not* get their own

table:


>

> This should have been:

> - Non-identifying one-to-many relationships without singlevalued


attributes


> do *not* get their own table:

>

> > - Non-identifying one-to-one relationships do *not* get their own table:

>

> This should have been:

> - Non-identifying one-to-one relationships without singlevalued attributes


> do *not* get their own table:

>

> Apologies for the inconvenience. As a final remark let me say that you

might


> consider to always map a relationship to a table, but in some cases (esp.


> identifying relationships) this gives you a redundant database scheme

where


> the relationship-table is simply the projection of the weak-entity table.


>

> -- Jan Hidders

>

>