Re: Does Codd's view of a relational database differ from that ofDate& Darwin? [M.Gittens]
From: mountain man <hobbit_at_southern_seaweed.com.op>
Date: Wed, 08 Jun 2005 02:49:20 GMT
Message-ID: <Q0tpe.7446$F7.3138_at_news-server.bigpond.net.au>
Date: Wed, 08 Jun 2005 02:49:20 GMT
Message-ID: <Q0tpe.7446$F7.3138_at_news-server.bigpond.net.au>
"Jon Heggland" <heggland_at_idi.ntnu.no> wrote in message
news:MPG.1d0fda8950d6126d989683_at_news.ntnu.no...
> In article <42a5b411$0$8714$ed2619ec_at_ptn-nntp-reader02.plus.net>,
> paul_at_test.com says...
>> And that all
>> aggregates should return NULL if any of the attributes in the aggregated
>> column are NULL. Surely if any of them are unknown, we must say that the
>> sum or average is unknown also?
>
> Yes, if NULL means unknown.
No. We can certainly sum or average those which are not null, and this is how the MS SQL handles nulls in aggregates: they are ignored, with the following message:
Warning: Null value is eliminated by an aggregate or other SET operation.
-- Pete Brown IT Managers & Engineers Falls Creek Australia www.mountainman.com.au/softwareReceived on Wed Jun 08 2005 - 04:49:20 CEST