Re: RMAN impact

True enough! Sort of. Except that a primary key implies an index. As does a foreign key. (Actually, unless you're crazy, many foreign keys will imply two indexes, one on each table.)

Conversely, no indexes means no primary keys. In my personal fantasy world, lack of Primary Keys means unemployed data modelers, but sadly the so-called real-world seems to be quite different... ;-)

A (logical) data model may not specify indexes, but if it's done correctly, it's going to imply at least one on every table. If the physical data model fails to embody those indexes, then somebody ought to be in trouble! ;-)

On 10/26/06, Jared Still <jkstill_at_gmail.com> wrote:
>
> On 10/25/06, Mark Brinsmead <pythianbrinsmead_at_gmail.com> wrote:
> >
> > than they ought to be. If you have the extraordinary bad luck of having
> > such outdated
> > blocks occur (for example) in a table with no indexes (you'd want to
> > talk to your data
> > modeler about that anyway) then you would have no (useful) way of
> > detecting this
> > corruption.
> >
> >
> Data Modelers don't design indexes.
>
> Unless the DA and DBA are the same person.
>
> --
> Jared Still
> Certifiable Oracle DBA and Part Time Perl Evangelist
>

-- 
Cheers,
-- Mark Brinsmead
   Senior DBA,
   The Pythian Group
   http://www.pythian.com/blogs

--
http://www.freelists.org/webpage/oracle-l