Question

The following program expects user input in the mixed fraction format ‘whole_numbernumerator/denominator’ and assigns values to respective variables.

#include<iostream>
using namespace std;

int main()
{
    int whole, numerator, denominator;

    cout << "Input format: i<space>n/d" << endl;

    cin >> whole;
    cin.ignore(1000, ' ');
    cin >> numerator;
    cin.ignore(1000, '/');
    cin >> denominator;

    cout << whole << endl;
    cout << numerator << endl;
    cout << denominator << endl;

    return 0;
}

Input1:
123 345/678
Output1:
123
345
678

Input2:
1111111111 1111111111/1111111111
Output2:
1111111111
1111111111
1111111111

Input3:
2222222222 2222222222/222222222
Output3:
2147483647
0
0

I haven’t been able to figure out why the program doesn’t work for Input3.

Answer 1

You’re overflowing the maximum value for a 32-bit integer (2^31-1 ~= 2.147b). Once this happens, cin doesn’t work properly until you clear the flag. You should be checking for errors, but a short-term solution is to make your number unsigned, or use a 64-bit one, like int64_t. You also don’t need to ignore the space, as cin will skip it by default.

You can implement something like what’s found here to ensure valid input, but it needs to be tailored to fit your specific input format. Perhaps encapsulating the three into a single type with an overloaded operator that inputs each with respect to the formatting would make the syntax fit more nicely, so you could replace age in the example with a MixedNumber object.

I would see something like this as a versatile method:

template <typename T> //any type will work
void getValidInput (T &var, std::string prompt = "Input: ") {
    while ((std::cout << prompt) && !(std::cin >> var)) { //if cin fails...
        std::cin.clear();                 //clear flag and discard bad input
        std::cin.ignore (std::numeric_limits<std::streamsize>::max(), '\n');
        std::cout << "Invalid input; please re-enter.\n"; //let the user know
    } 
}

Then you could just have your program as follows:

struct MixedNumber { //a data structure, so it's like using plain variables
    int64_t whole; 
    int64_t numerator;
    int64_t denominator;
};

std::istream &operator>> (std::istream &in, MixedNumber &num) { //so cin works
    in >> num.whole >> num.numerator;
    in.ignore(); //yes, you could enforce the format a bit more
    in >> num.denominator;
    return in;
}

int main() {
    MixedNumber num; //easy to "make" a mixed number, a constructor works well too
    getValidInput (num, "Input format: i<space>n/d: "); 
    std::cout << num.whole << '\n' << num.numerator << '\n' << num.denominator;
}