Help with float validation

[ashwyn]

Hello.
I'm writing a program and I need it to accept input from the user in main, and then call a function that will validate that input as a legal float data type.

I've tried to accept user input as a string, then iterate through the string to make sure all characters are acceptable float characters.
But I'm realizing I don't know how to do it very well, and need help figuring out a good method for input validation.

Here's my code. It isn't finished, and doesn't work well. But I feel pretty stuck with my knowledge.

#include <iostream>
#include <string>
#include <cmath>
#include <cstdlib>


using namespace std;


float valid_float(string str){

   int dot = 0;
   float val;

   for(int j = 0; j < str.length(); j++){      // This is a lame attempt to make sure there is only a single '.' in the answer, and no more.
   
      if(str[j] == 46){
      
	 dot += 1;
      }
      if(dot > 1){
      
	return 0000000000000001;
      }
   
   }
   for(int i = 0; i < str.length(); i++){
   
      if(!(str[i] >= 48 && str[i] <= 57)){
      
	                                 // Cast string as a float
      }
      
      
      
      


   
   }


}


int main(){
float x;
string input;

do{
cout << "\nPlease enter a valid float: " << endl;
cin >> input;
x = valid_float(input);
}while(x == 0000000000000001);            // This is supposed to catch errors when a user enters more than one '.'
return 0;
}

Thank you very much in advance for help on this.

[tpb]

Don't use numeric codes for characters. Just say '.' and '0', etc.

Do you think those leading zeroes are doing something? It's still just 1.
If you want a special value to return if you weren't able to convert the string you could return NaN (numeric_limits<double>::quiet_NaN()). Test it with isnan().

Anyway, if you haven't already tried to do this exercise with an int, then I'd start there first.

[Satan]

Can you use functions? Then you can rely on something like stof() to directly convert the string to floating.

Otherwise you have to convert the string to integer and then divide by 10 for every decimal place to the right of the floating point.

[JLBorges]

// if valid float, place the result of the conversion in value and return true
// return false if the entire string could not be interpreted as float
bool valid_float( std::string str, float& value )
{
    // remove trailing white space, if any
    while( !str.empty() && std::isspace( str.back() ) ) str.pop_back() ;

    // https://en.cppreference.com/w/cpp/string/basic_string/stof
    std::size_t pos = 0 ;
    try { value = std::stof( str, &pos ) ; } // try to interpret the string as float
    catch( const std::exception& ) { return false ; } // conversion failed: not a valid float

    return pos == str.size() ; // valid float if all characters in the string have been consumed
}

http://coliru.stacked-crooked.com/a/239e0ce07421e8c1

[tpb]

I assume that it's an exercise and therefore he can't use functions. Otherwise I wouldn't given the obvious answer. Obviously.

[lastchance]

and then call a function that will validate that input as a legal float data type

#include <iostream>
#include <sstream>
#include <string>
using namespace std;

bool isValidFloat( string s, float &f )
{
   char junk;
   stringstream ss( s );
   return ( ss >> f && !(ss >> junk) );
}

int main()
{
   float f;
   string tests[] = { "1", "2.0", "3.5e4", "3.5e400", "  -3.145  ", "2.5  3.6", "abc" };

   for ( string s : tests )
   {
      if ( isValidFloat( s, f ) ) cout << "\"" << s << "\" is the valid float " << f << '\n';
      else                        cout << "\"" << s << "\" is not a valid float\n";
   }
}

"1" is the valid float 1
"2.0" is the valid float 2
"3.5e4" is the valid float 35000
"3.5e400" is not a valid float
"  -3.145  " is the valid float -3.145
"2.5  3.6" is not a valid float
"abc" is not a valid float