Re: min(); never no_data_found
Date: Thu, 24 Dec 2009 10:23:16 -0800 (PST)
On Dec 24, 2:02 am, Jeremy <jeremy0..._at_gmail.com> wrote:
> In article <799a158c-7229-4232-82a0-264a4a6303c0
> _at_d4g2000pra.googlegroups.com>, joel-ga..._at_home.com says...>
> > On Dec 23, 7:10 am, Jeremy <jeremy0..._at_gmail.com> wrote:
> > > In article <79bb2ee5-f28a-43b9-af1e-4597012495f0
> > > _at_a32g2000yqm.googlegroups.com>, Mark.Powe..._at_hp.com says...
> > > > Cate, why not test for a NULL value being returned and if so the RAISE
> > > > NO_DATA_FOUND?
> > > What's the opinion of this ng's contributors on the advisability of
> > > raising "standard" error conditions when the underlying reasons for
> > > raising that error are different?
> > > To me it seems like a practice not to be recommended - potential for
> > > misleading people unfamiliar with the code later on in maintenance mode.
> Not sure why you posted that - my question is simply is it bad practice
> to raise an exception that means something specific when the underlying
> condition found is something diffferent? As in the example here where
> someone is finding that a value is NULL but wishes to treat that as
> jeremy- Hide quoted text -
> - Show quoted text -
Not so much bad practice as bad execution. The idea of trapping for an error means you are handling an error. If you are handling the error wrong that would be a bug. The example of handling null as no_data_found... I would suggest googling on 3 and 4 value null debates that have been going on since forever. The answer, both for your question and example, is undefined.
-- _at_home.com is bogus. http://www.signonsandiego.com/news/2009/dec/24/doomed-china-restaurant-hires-live-in-protester/Received on Thu Dec 24 2009 - 12:23:16 CST