Re: Natural keys vs Aritficial Keys

From: paul c <>
Date: Mon, 18 May 2009 17:16:30 GMT
Message-ID: <OJgQl.28538$PH1.21741_at_edtnps82>

Bob Badour wrote:

> Roy Hann wrote:

>> paul c wrote:
>>> To most people, "natural" means "physical"
>> Really?! I can't possibly say you are wrong about that because I
>> wouldn't know, but you do surprise me.
>>> and ends up being quite unnatural for some purposes.. When inventory
>>> or anything else "moves", sometimes "logical" locations are more
>>> convenient, allowing the replacement of one row to indicate the
>>> transfer or re-assignment of many "things".
>> That seems very implausible in the real world. That is a logical model
>> in which if you move one thing, you must--infallibly--move all the
>> other things that were there too. That is, it is a world in which it is
>> impossible not to also move all the other things. Such situations do
>> exist, I grant you, but they are the exception not the rule. But no
>> one here (AFAIK) has ever said synthetic/surrogate keys are never
>> useful.
> And no synthetic key will solve the problem that someone was stupid 
> enough to fill a warehouse with no internal subdivision and no way to 
> identify where in the warehouse inventory was put. (In the example 
> cited, the warehouse being a huge gravel field.)

If management is too stupid to sub-divide the yard, that doesn't prevent a db from assigning a value that indicates pieces are adjacent, based on sequence of initial movement, or whatever. Just like nothing prevents a system from moving all employees in a department to a different department by replacing a single row. Received on Mon May 18 2009 - 19:16:30 CEST

Original text of this message