<< Also note that blank and empty are not null values, they are zero length strings or the equivalent in numbers. >>
Except that there isn't any empty value for numerical data -- 0 is a number, not emptiness. I understand the logic of not allowing NULLs to be compared to each other but I think in the absence of any consistent way (and, in many cases, any way at all) to specify "known emptiness" that NULL logic is a flaw in the design of SQL. Somewhere in the list of normal forms is a requirement to ensure that there are no columns in your database that allow NULL values -- obviously that's an extreme level that no one ever goes to. As evidence for NULL logic being a flaw consider how often this subject comes up here, how hard it is for people to understand it, the contortions we have to go through to program around it (eg, Jame's original example in this thread) and, finally, the fact that most database languages have an "out" like R:Base's EQNULL setting. -- Larry
