On Fri, 23 Feb 2001 08:57:26 +0200, Gerhardus Geldenhuis wrote:

>Thanks for all the replies.
>I like to try to stay with the current way of doing things.
>
>However I have no idea what surrogate keys are. Unique ID's
>well the name tells it all. I also presume that unique ID's
>are char fields or can it be integers ?

What lady Borrie, Nando, et all, call surrogate keys are
(in general) sequential integer keys with no business meaning. They exist
to support relational constraints, and using then you not
cascade updates when business columns are altered;
it's good, because this is got without mess the data.
This 'technique' have its downsides too, but in IT -- and
life -- nothing is perfect. ;-)

>Would Database Design for mere mortals be a good book
>to teach me some/all of these priniciples ?

There are lots of them, ones more easy, other more difficult.
Wow, I'm seeing the book in my mind and the name I can't
remember!
But it remember me:

[Old memory on - a little off-topic]

The best one I got is from a author from here (Brazil)
and the book is a part of a 4 tomes collection about.... Clipper!
(Believe or not). The fourth tome gets in database
design in a so good explained that I never made a normalisation
mistake after reading...

[Old memory off]

>
>Nando Dessena wrote:
>
>> Gerhardus,
>>
>>
>>> You mentioned that this method is maybe obsolete. Can you tell me what
>>> the new way of doing this is.
>>
>>
>> Surrogate keys and unique IDs; no composite keys.
>>
>>
>>> Or the correct way in firebird/ib to do
>>> this without performance degradence.
>>
>>
>> With IB it is more healthy to avoid defining foreign keys on low
>> selectivity columns; it depends on the nature of your m/d relationship.
>> I'd tend to use one in your case.
>> Ciao
>
>
>
>To unsubscribe from this group, send an email to:
>ib-support-unsubscribe@egroups.com
>
>
>
>Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>

[]s Fabricio
Systems Developer