Marc, I still advocate the use of Missing Data Values in place of your NULLs. I'd work with the users to see what types/classes of missing data they might need. It could be as simple as "N/A", or "Not Available" (watch out, though, because to some folks "N/A" means "Not Applicable", which is subtly but discretely different from "Not Available).
Again, these are actual values, not symbolic for NULL, i.e. it is not R>SET NULL 'N/A' In the case of INTEGER or any kind of numeric value, it can be trickier. As I said previously, I used negative integers for different types of missing data when I worked in the area of survey research as all the user-supplied responses were positive integers or text. Once defined, you could even use referential integrity/rules enforcement, as well as some sort of lookup help for the user, to require use of such values at entry/edit which, of course, could eliminate the existence of most/all nulls. Another $0.02, Steve in Memphis -----Original Message----- From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of Marc Sent: Monday, April 14, 2008 7:40am 07:40 To: RBASE-L Mailing List Subject: [RBASE-L] - Re: EQNULL ON or OFF Thanks James I guess I am trying to look at this from the users stand point. To the user the field either had data or no data, I guess I am having a hard time thinking of a situation how having EQNULL ON will cause me a problem. thanks for the education Marc ----- Original Message ----- From: "James Bentley" <[EMAIL PROTECTED]> To: "RBASE-L Mailing List" <[email protected]> Sent: Friday, April 11, 2008 2:16 PM Subject: [RBASE-L] - Re: EQNULL ON or OFF > Marc, > > The three way logic introduced normal boolean logic by having > NULL as a valid value is enough to make any one's mind spin. > Hence, my rule not to futz with the EQNULL setting. I always > leave it OFF and muddle through my tests until I get a > satisfactory answer. It is to easy to turn it on in one section > and forget to turn it off while having another section blow up > in your face. > > My feeling is that any syntax/switch that allows you to > circumvent standard SQL is a time bomb waiting to happen. When > several programmes modify the code using non-standard SQL syntax > who knows what will happen. > > > --- Marc <[EMAIL PROTECTED]> wrote: > >> This always reminds me of Who's on First, it makes your head >> spin. >> >> But to my way of thinking and what I think would make sense to >> users >> >> Hello is only = to Hello therefore <> to anything else >> >> so with EQNULL ON or OFF the Pause statement should fire >> because >> no matter what Hello can not be = to a blank field or a null >> or empty field >> Hello can only be = to Hello >> >> The same for <> Hello is = to only Hello therefore <> to >> everything else. >> >> Am I nuts or just not getting this. Maybe my RBrain is set >> Off? >> >> Marc >> >> >> >> >> ----- Original Message ----- >> From: [EMAIL PROTECTED] >> To: RBASE-L Mailing List >> Sent: Friday, April 11, 2008 8:31 AM >> Subject: [RBASE-L] - Re: EQNULL ON or OFF >> >> >> Marc: >> >> For you, the danger would be to set it OFF right now without >> checking your programs. If your programs were done with the >> assumption that eqnull was set ON, then you better not change >> the setting. Here's the main difference: >> >> set var vtext1 = 'hello', vtext2 text = null >> if vtext1 <> .vtext2 then >> pause 2 using 'they are not equal, so do something' >> endif >> >> if eqnull ON, then the pause would evaluate because it is >> able to compare a null to a value. If you eqnull is OFF >> (which I believe the majority of us do), then a null cannot be >> compared with anything and the pause would NOT evaluate. >> >> Karen >> >> >> >> >> Marc, >> >> When EQNULL is on, R:BASE doesn't distinguish a null from >> a zero, so any average calculations are screwy. >> >> Bill >> >> On Fri, Apr 11, 2008 at 8:41 PM, Marc >> <[EMAIL PROTECTED]> wrote: >> >> >> After reading the other thread I got nervous about my >> EQNULL setting. >> I have it set to ON, so now I am worried. I had a >> problem sometime back >> and setting EQNULL ON fixed that problem so I just leave >> it on. >> >> What are the dangers with it ON? >> >> thanks >> Marc >> >> >> >> >> >> > > > Jim Bentley > American Celiac Society > [EMAIL PROTECTED] > tel: 1-504-737-3293 > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > >
