Re: Null in subquery returns no records

From: Dawn M. Wolthuis <dwolt_at_tincat-group.com>
Date: Sat, 7 Feb 2004 08:26:49 -0600
Message-ID: <c02sjp$q4n$1_at_news.netins.net>


"Jonathan Leffler" <jleffler_at_earthlink.net> wrote in message news:z%0Vb.17554$uM2.2847_at_newsread1.news.pas.earthlink.net...
> Mike MacSween wrote:
> > New testing rule for Mike - always test with null values.
>
> Alternative rule - always ensure the database stores no nulls!
>
> Not necessarily easy - but a good discipline.
>
<snip>
And perhaps the best option of all -- use one of the databases that employs a two-valued logic, with NULLs treated as null sets -- much cleaner and clearer. Examples are jBASE from www.jbase.com , Revelation from www.revelation.com and the U2 databases (UniData and UniVerse) from www.ibm.com (then do a search). You would not be alone as there are over 100,000 companies that use these products (could be a low estimate).

I'm sure there are others as well that have avoided many of the SQL and RDBMS complications. There is nothing holy nor scientifically more correct about the RDBMS model than many other possibilities. If you have an opportunity to free yourself of the notion that an RDBMS is required, you are likely to find a more productive environment elsewhere.

Just thought I'd toss in my two cents even though most people who share this opinion keep it quiet. --dawn Received on Sat Feb 07 2004 - 15:26:49 CET

Original text of this message