C++ Convert String to Int

Previously discussed Convert Char to Int in C++, in which we converted a single character to an integer. In this article we will build upon that idea by converting multiple characters to an integer.

C++ Standard Library

If want to use the built-in functions, the C++ standard has a couple built-in functions for converting strings to an integer. Namely there is atoi and std::stoi.

atoi

atoi is part of “C Standard General Utilities Library in the header <cstdlib>.

#include <cstdlib> // atoi
#include <cstdio> // printf

int main()
{
    const char* buffer = "123456";
    int num;

    // int atoi(const char *buffer);
    num = atoi(buffer);

    if (num == 0)
    {
        // error
        printf("Could not convert number");
    }
    else
    {
        // output value
        printf("Number is %d\n", num);
    }

    return 0;
}

Notes: atoi returns 0 on error, which can be confusing if the input was intended to be zero. If the converted value is outside the range of int then it is undefined behavior.

stoi

stoi is part standard name space and is in the C++ <string> header. stoi has the added benefit of being able to convert different bases.

#include <string> // std::string
#include <iostream> // std::cout

int main()
{
	using namespace std;

	string str1 = "12345"; // decimal number (base 10)
	string str2 = "  123 abcd"; // can be part of string, but has to part of the beginning
	string str3 = "c0ffee"; // base 16
	string str4 = "0xff"; // can auto detect

	int num1 = stoi(str1); // default is Idx = nullptr, Base = 10
	int num2 = stoi(str2, nullptr, 10);
	int num3 = stoi(str3, nullptr, 16);
	int num4 = stoi(str4, nullptr, 0); // can auto detect

	cout << str1 << " -> " << num1 << endl;
	cout << str2 << " -> " << num2 << endl;
	cout << str3 << " -> " << num3 << endl;
	cout << str4 << " -> " << num4 << endl;

	return 0;
}

The difference between atoi and stoi

Unless C strings must be used, then it is advisable to use stoi instead. One pitfall of atoi is that returns 0 on error, which can be valid input to the function. If stoi can’t converted the buffer into an integer it throws an exception (either invalid_argument or out_of_range exception).

We can see that C++ stoi has more safety features than C atoi, which comes at the cost of dealing with exceptions.

Convert multiple characters to a single int

We can iterate on the idea of converting one character into converting multiple characters into an integer. The interesting part of this algorithm is the computation of the result. It is done by 10 * result + ch - '0'. First, the ones digit is computed by ch - '0'. Next, the result is shifted by 10. Then the new digit added.

Initialization:
result = 0

Iteration 1:
digit = 1
result = 10 * result + digit = 1

Iteration 2:
digit = 2
result = 10 * result + digit = 10 + 2 = 12

Iteration 3:
digit = 3
result = 10 * result + digit = 120 + 3 = 123

Iteration 4:
digit = 4
result = 10 * result + digit = 1230 + 4 = 1234 // final result

We can represent this idea in code:

unsigned int my_stoi(string s)
{
    unsigned int result = 0;

    for (char ch : s)
    {
        if (isdigit(ch))
        {
            int digit = ch - '0';
            result = 10 * result;
            result = result + digit;
        }
    }

    return result;
}

Note that this is only a sketch of what the algorithm should be. For example, the down side to this is we can have a string that is longer than what can fit into an integer (i.e., out of range). We can use a larger integer type, such as unt64_t but this too has a limitation on how many digits can be converted. The second assumption it makes is that the number is positive. Third assumption it makes that there is no extra whitespace around or between the digits (in this algorithm it ignores it). This exercise can be left with the reader to