is a numeric value ECMA-404

[andresdt]

Hello
I share with you this function that works perfectly for me. Determine if the value of a variable is a numerical value according to the ECMA-404 standard.



The function code:

Code: Select all

function isNumericValue pValue
    local tFirstChar
    put char 1 of pValue into tFirstChar
    if tFirstChar is 0 and char 2 of pValue is not "."  or tFirstChar is not in "-0123456789"
    then return false

    # Scientific notifications
    if "e" is in pValue then
        set the itemdel to "e"
        if the number of items of pValue is 2 and char 1 of pValue is not  in".e" and char -1 of pValue is not  in".e" then
            if char 1 of item 2 of pValue is in "+-" then
                delete char 1 of item 2 of pValue
            end if
            put the item 1 of pValue & the item 2 of pValue into pValue
        else
            return false
        end if
    end if

    split pValue by "."
    if the number of elements of pValue > 2 then return false
    if pValue[ 2 ] is empty then  return pValue[ 1 ] is an integer
    return pValue[ 1 ] is an integer and pValue[ 2 ] is an integer
end isNumericValue

[andresdt]

I ran into a bug in the function code.
Well, since when evaluating the number 0 it returned false.
fixed

Code: Select all

function isNumericValue pValue
    local tFirstChar
    put char 1 of pValue into tFirstChar
    if pValue is not an number then return false
    put char 1 of pValue into tFirstChar
    if tFirstChar is 0 and char 2 of pValue is not "." and  the length of pValue is not 1 or tFirstChar is not in "-0123456789"
    then return false
    
    # Scientific notifications
    if "e" is in pValue then
        set the itemdel to "e"
        if the number of items of pValue is 2 and char 1 of pValue is not  in".e" and char -1 of pValue is not  in".e" then
            if char 1 of item 2 of pValue is in "+-" then
                delete char 1 of item 2 of pValue
            end if
            put the item 1 of pValue & the item 2 of pValue into pValue
        else
            return false
        end if
    end if
    
    split pValue by "."
    if the number of elements of pValue > 2 then return false
    if pValue[ 2 ] is empty then  return pValue[ 1 ] is an integer
    return pValue[ 1 ] is an integer and pValue[ 2 ] is an integer
end isNumericValue 

[andresdt]

I was reviewing the function and I noticed that it could be optimized a lot. Since if I use the LiveCode function (value is a number) I would only have to worry about the numbers that start with 0 and are not 0 or 0. # and the LiveCode constants that are numbers and other strange ones like infinity. So this is how the new version looks it is considerably smaller and faster.

Code: Select all

function _isNumericValue pValue 
    if pValue is not a number then return false
    if char 1 of pValue is 0 and char 2 of pValue is not "." and the length of pValue is not 1 or char 1 of pValue is not in "-0123456789"
    then return false
    return true
end _isNumericValue

[Thierry]

....
So this is how the new version looks it is considerably smaller and faster.

Code: Select all

function _isNumericValue pValue 
    if pValue is not a number then return false
    if char 1 of pValue is 0 and char 2 of pValue is not "." and the length of pValue is not 1 or char 1 of pValue is not in "-0123456789"
    then return false
    return true
end _isNumericValue

or a bit smaller and probably easier to read?

Code: Select all

function _isNumericValue pValue 
   if pValue is not a number then return false
   return char 1 of pValue is 0 \
         and char 2 of pValue is not "." \
         and the length of pValue is not 1 \
         or char 1 of pValue is not in "-0123456789"
end _isNumericValue

Kind regards,

Thierry

[andresdt]

Thanks @Thierry.
In its version, for it to work correctly you must put everything that is returned in the second line between parentheses. Then evaluate whether everything in parentheses is false. Since like this , the opposite of what we want to obtain returns us. (For example for 2 it returns false and 002 returns true.)
In this way we can also put the first (if) on this line and it is even smaller. Also, it is a few milliseconds faster.
Thanks again. This is what it would look like:

Code: Select all

function _isNumericValue pValue 
   return pValue is a number and (char 1 of pValue is 0 \
             and char 2 of pValue is not "." \
             and the length of pValue is not 1 \
             or char 1 of pValue is not in "-0123456789")  is false
end _isNumericValue

[Thierry]

Hi,

in fact, I didn't check the validity of your function; I did trust you

just rewriting your code a tiny bit....

this time, I drop the last part "is false" and add not before (

Code: Select all

function _isNumericValue pValue 
   return pValue is a number \
             and not ( char 1 of pValue is 0 \
             and char 2 of pValue is not "." \
             and the length of pValue is not 1 \
             or char 1 of pValue is not in "-0123456789")
end _isNumericValue

and I don't think you'll win few milliseconds,
but the code is certainly more beautiful.

So, what's next

Thierry