Re: Insert with response
From: Hugo Kornelis (hugo_at_pe_NO_rFact.in_SPAM_fo)
Date: 11/07/04
- Next message: Hugo Kornelis: "Re: Please help SQL"
- Previous message: David Browne: "Re: Please help SQL"
- In reply to: Louis Davidson: "Re: Insert with response"
- Next in thread: Louis Davidson: "Re: Insert with response"
- Reply: Louis Davidson: "Re: Insert with response"
- Messages sorted by: [ date ] [ thread ]
Date: Mon, 08 Nov 2004 00:41:54 +0100
On Sun, 7 Nov 2004 15:29:13 -0600, Louis Davidson wrote:
>> I can't speak for Louis, but I'll tell you that this WILL work. In fact,
>
>True :)
>
>> it will even work when the same applicant tries to register again - if you
>> defaine only an identify column and nothing else as PRIMARY KEY or UNIQUE,
>> you'll never have duplicate key violations again - not even if you really
>> should have them.
>
>The lack of error messages does not mean that it works. The rest of your
>post seems to contradict this point.
Hi Louis,
If you'd seen my face when I wrote this message, you'd have known that the
"it will work" remark was sarcastic - I tried to put across the point that
this design will "work" (i.e. accept entered data) even when it soould not
"work" (i.e. it should reject the data for being duplicate data).
> Identities have limitations which should be understood, but they
>work very nicely as surrogate keys.
My point exacly. Unfortunately, the world is still littered with a
depressingly large population of database designers who will just happily
stick an identity column on a semi-random collection of columns and think
that they now have a good design. <sigh>
Best, Hugo
-- (Remove _NO_ and _SPAM_ to get my e-mail address)
- Next message: Hugo Kornelis: "Re: Please help SQL"
- Previous message: David Browne: "Re: Please help SQL"
- In reply to: Louis Davidson: "Re: Insert with response"
- Next in thread: Louis Davidson: "Re: Insert with response"
- Reply: Louis Davidson: "Re: Insert with response"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|