Validate Whole Number or Zero with Function

[Jer]

I am trying to use a function to validate numeric input. This method, not in a function,
was recently posted as a solution in this forum.
Can you tell me why the function below does not communicate the result, which would be FALSE
in this example, after the function exits? After the call line, the variable isNumb when echoed is blank.

Code: Select all

@Echo Off
setlocal enabledelayedexpansion

Set variable=12b

Call:checkNumber %variable% isNumb
Echo %variable% -- is it zero or a whole number? !isNumb!

GoTo:eof

:checkNumber
setlocal enabledelayedexpansion
Set isNumb=TRUE

For /f "delims=0123456789" %%a In ("%1") Do Set isNumb=FALSE

Echo Line 17, testing...isNumb: !isNumb!
endlocal & Set %~2=!isNumb!
Exit /b

endlocal 

[julesverne]

I did this, and I got FALSE as the variable: I switched your goto:eof and your exit /b around. I added the pause so you could see the result.

Code: Select all

@Echo Off
setlocal enabledelayedexpansion

Set variable=12b

Call:checkNumber %variable% isNumb
Echo %variable% -- is it zero or a whole number? !isNumb!
pause
Exit /b

:checkNumber
setlocal enabledelayedexpansion
Set isNumb=TRUE

For /f "delims=0123456789" %%a In ("%1") Do Set isNumb=FALSE

Echo Line 17, testing...isNumb: !isNumb!
endlocal & Set %~2=!isNumb!
GoTo:eof

endlocal 

[penpen]

This is the issue:

Code: Select all

endlocal & Set %~2=!isNumb!

The endlocal (in the function :checkNumber) restores the old environment, so the variable value of isNumb is restored (empty in this case).
I think you wanted to use %isNumb%, in this case, as the value is set before this line is executed:

Code: Select all

@Echo Off
setlocal enabledelayedexpansion
Set variable=12b

Call:checkNumber %variable% isNumb
Echo %variable% -- is it zero or a whole number? !isNumb!

endlocal 
GoTo:eof

:checkNumber
setlocal enabledelayedexpansion
Set isNumb=TRUE

For /f "delims=0123456789" %%a In ("%1") Do Set isNumb=FALSE

Echo Line 17, testing...isNumb: !isNumb!
endlocal & Set %~2=%isNumb%
Exit /b

penpen

[Aacini]

May I suggest a different method to validate the input? You may use the subroutine described at this post that was developed precisely for this task. This way, the user is allowed to enter just digits, so the test is not required at all! That subroutine may be modified in order to just read a number of N digits, so the code becomes shorter...

Antonio

[rojo]

I am trying to use a function to validate numeric input. This method, not in a function,
was recently posted as a solution in this forum.
Can you tell me why the function below does not communicate the result, which would be FALSE
in this example, after the function exits? After the call line, the variable isNumb when echoed is blank.

I agree with penpen. Whenever you endlocal & set "var=!var!" delayed expansion causes heartache and misery. Use endlocal & set "var=%var%" or endlocal & set "var=%%I" instead.

Since you're basically asking your subroutine to return a boolean value, you could have it set %ERRORLEVEL% and evaluate it with conditional execution like this:

Code: Select all

@echo off
setlocal

for %%I in (12345 123.3 0 abcde "12") do (
   call :isInt %%I && (
      echo %%I is an integer.
   ) || (
      echo %%I is not an integer.
   )
)

exit /b 0

:isInt <str>
for /f "delims=0123456789" %%a in ("%1") do exit /b 1
exit /b 0

Result:

Code: Select all

C:\Users\me\Desktop>test
12345 is an integer.
123.3 is not an integer.
0 is an integer.
abcde is not an integer.
"12" is not an integer.

[Compo]

What's wrong with this?

Code: Select all

Echo(%variable%|FindStr "[^0-9]">Nul&&(Echo(Non integer)||(Echo(Integer)

[rojo]

What's wrong with this?

Code: Select all

Echo(%variable%|FindStr "[^0-9]">Nul&&(Echo(Non integer)||(Echo(Integer)

Internal functions are generally faster than executables. Your solution is simpler, but less efficient.

[Squashman]

What's wrong with this?

Code: Select all

Echo(%variable%|FindStr "[^0-9]">Nul&&(Echo(Non integer)||(Echo(Integer)

Yep. Just another way to skin the cat. I know we used that one just a few weeks ago. Had to explain inverse class to the user.

[Jer]

Thank you all for your input to this post. I am going with penpen's fix to my batch code, using %var% instead of !var!
notation to capture the result of the function.

This is for a hobby-type batch file application, so in my early stages of learning I will stick with the simpler solution.
Jerry