Scott Ellsworth <scot..lodar.com> wrote:
>
> On May 3, 2004, at 10:24 AM, Andrus Adamchik wrote:
>
> > I guess nobody cared to implement this, since having a separate join
> > table
> > for each n:m is arguably cleaner design and makes db data maintenance
> > easier. But I'd like to hear arguments in favor of such generic join
> > table
> > design.
Not sure if this counts, but I have a ChangeLog table that logs all write
operations to other database tables.
One of the fields is "FOREIGN_KEY" (of the changed record) which can be the
foreign key to any other table in the database.
I handle it by defining lots of relationships all using this same field.
This archive was generated by hypermail 2.0.0 : Mon May 03 2004 - 14:02:08 EDT