Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Would be really nice if...

Re: Would be really nice if...

From: Serge Rielau <srielau_at_ca.ibm.com>
Date: Thu, 24 Feb 2005 07:20:31 -0500
Message-ID: <385v3eF5iovfqU1@individual.net>


Haximus wrote:
>
> As with any 'feature' usage is optional, you are not OBLIGATED under threat
> of a gun to your head to use it, if it offends your sensibilities or
> personal tastes. Morgan's example is really stretching it if that's the
> only thing he could come with... that there are considerations when
> modifying table structures in any situation is an IMPLIED NO-BRAINER. Using
> a feature suggests that you know what the hell you are doing with it in the
> first place... and understanding it's effects.
>
> I'm not sure what the big conceptual problem is with it, it's simply an
> alternate method to specify a column select list... that's about the grand
> summation of it.

I challenge the fact that users undertsand what they are doing. In my experience, and across vendor borders, thie is simply not true. There are plenty of voices claiming that SELECT * was bad idea to begin with. It causes endless grief w.r.t. schema evolution.

Having said that I don't really think that an EXCLUDE option makes things any worse than they already are.
But while you're at it the next step will be an EXCLUDE syntax on INSERT to avoid having to specify a column list when dealing with columns that are not updatable... and so on.
It's a bit of a complex-instruction set vs. reduced instruction set problem. Is it worth bloating the language further because of syntactic convenience given the existence of modern tooling?

Cheers
Serge

-- 
Serge Rielau
DB2 SQL Compiler Development
IBM Toronto Lab
Received on Thu Feb 24 2005 - 06:20:31 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US