Re: So what's null then if it's not nothing?
Date: Thu, 24 Nov 2005 07:35:52 GMT
Message-ID: <s3ehf.2601$ea6.1877_at_news-server.bigpond.net.au>
"FrankHamersley" <FrankHamersleyZat_at_hotmail.com> wrote in message
news:EYXgf.1705$ea6.525_at_news-server.bigpond.net.au...
> mountain man wrote:
>> "paul c" <toledobythesea_at_oohay.ac> wrote in message >> news:CIQgf.578722$tl2.426979_at_pd7tw3no... >>>mountain man wrote: >>>>"paul c" <toledobythesea_at_oohay.ac> wrote in message >>>>>Frank Hamersley wrote: >>>>>>JOG wrote: >>>>>>>michael_at_preece.net wrote:
> [..]
>> >> The technical implementation of the handling of null values >> within the different RDBMS and DBMS software vendors >> such as Microsoft, Oracle, IBM, etc are diverse, as are >> the various kinds of breakfast cereals. >
> Do you know of a reference that explains these variations succinctly - I
> must admit my breadth of vendor quotient is quite low.
This may have some comparisons in some section: Are SQL Server, DB2, and Oracle really relational? http://www.handels.gu.se/epc/archive/00002948/
> Until now I had assumed most (if not all) performed the 3VL fandango in
> the same direction around the dance floor - especially the (NULL == NULL)
> is FALSE outcome.
They have a lot of common elements, but you'll find a set of divergent properties as well.
> I had (previously?) thought the angst in this issue was more driven by ppl
> being aggravated by the "apparently" non-sensical outcomes when NULL
> values appear in join elements and/or procedural code and lead to
> unexpected results (bugs).
I look at this sort of stuff as data integrity issues, and a fact of life to be resolved by the IT crew at any database site.
Pete Brown
www.mountainman.com.au
Received on Thu Nov 24 2005 - 08:35:52 CET