## 14891Re: [Clip] isnumber function

Expand Messages
• Mar 9, 2006
• 0 Attachment
> I think the problem is with the definition of IsNumber in Help. It
> takes an argument of "Value" and then refers to "Value" as the
> "specified text". A better definition might be:
>
> \$IsNumber*("Str")\$ *Returns 1 if Str contains only digits 0 through 9,
> and 0 if it does not.
>
> This explains what you observe. The fact that these numbers are
> integers is just a coincidence. This definition also corresponds more
> closely to the definition of IsAlpha.
>
> It also might have avoided this whole thread. :-)

Your definition would have indeed been better than the one provided ...
but I think it is more accurate to say integers (defined as all whole
numbers positive AND NEGATIVE and zero:
http://en.wikipedia.org/wiki/Integer ). Your definition forgot the
negative sign ... that is not a 0-9 character ... and can only appear
immediately preceeding the number:
^!Info [C] minus 5 -- ^\$IsNumber("-5")\$
^!Info [C] 5 minus -- ^\$IsNumber("5-")\$
^!Info [C] 5 minus 5 -- ^\$IsNumber("5-5")\$

This tread would have still been necessary because what I needed to do
was test if something was a number in this format ([brackets mean
optional] and ... ellipses means one or more):
[#...][,][##]#[.][#...]

As isnumber didn't cover my situation ... I had to come up with a way to
do it.

Is is possible to test if something is a "number" with a regex?

Rules:
1. either one or no decimal point, but there can be any number of proper
characters before and after the decimal
2. comma's are permitted but not required, but may only be present with
three proper characters to the right of each comma and left of the
decimal (or implied decimal if it is an integer)
3. proper characters are 0-9, an no other characters are permitted
except for rule 1 (decimal point) and rule 2 (properly placed commas)

I wrote a clip function to do it, but if there is a regex that would do
it quicker ...

I just did this:
a+b+c+d+e+t+x+y=0
But the fact that those letters for some reason have a numeric value of
zero when used in calculations DOES NOT make them numbers per isnumber
when not used in calculations, as each of the following tests negative:
^!Info [C]abcdetxy --^\$IsNumber("abcdetxy")\$
^!Info [C]a --^\$IsNumber("a")\$
^!Info [C]aa --^\$IsNumber("aa")\$
^!Info [C]e --^\$IsNumber("e")\$

But when we use calc, we are all over the place ...
^!Info [C]abcdetxy --^\$IsNumber("^\$Calc(abcdetxy)\$")\$
^!Info [C]a --^\$IsNumber("^\$Calc(a)\$")\$
^!Info [C]aa --^\$IsNumber("^\$Calc(aa)\$")\$
^!Info [C]a*a --^\$IsNumber("^\$Calc(a*a)\$")\$
^!Info [C]e --^\$IsNumber("^\$Calc(e)\$")\$

Anyway, I am enjoying the tread and hope deep in my heart that this bug
just might be the catalyst to get Eric coding either an update or better
yet, a version 5. As I said a few months back though, I am not sure a
version 5 is really going to happen and nobody has told me I am wrong by
giving me any milestones, targets or goals. I'd settle for 4.96 with
this bug fixed and isnumber better defined in help ;-)
• Show all 22 messages in this topic