CODEWITHSUNDEEP
c++visual-studio-codec++14
TCat
TCat

Reputation: 13

Why is the following variable assignment not working correctly?

I tried to convert users input from string to int by casting each character to int and multiply by its corresponding exponential. Everything works correctly till the point I need to assign the result value to a int variable. It always one less than the correct value when the assigned value is between 100 and 1000. i.e. int k=0; k+=100; But printing k shows 99. But k+=1000 works fine.

The solution I found is to change the order of multiplicand and multiplier, but the cause to this situation is still unknown.

string k_input;
cin >> k_input;
int k = 0;
for (int i = 0; i < k_input.length(); i++)
{
    cout << (k_input[i] - '0') << " " << pow(10, (k_input.length() - i) - 1) << endl;

    k += (k_input[i] - '0') * pow(10, k_input.length() - i - 1);
}

i.e. By inputting 123, output should be 123, but it shows 122 instead.

And inputting 1221 shows 1221 as expected.

Upvotes: 1

Views: 350

Answers (1)

Adrian Mole
Adrian Mole

Reputation: 51825

The problem, which may very well be compiler- and/or platform-specific, is almost certainly down to 'rounding' errors when converting the double output of your calculation to an int value. You can fix this by using the lround function (defined in <cmath>):

k += lround((k_input[i] - '0') * pow(10, k_input.length() - i - 1));

Incidentally, turning on compiler warnings will highlight issues like this! For your code, in Visual Studio, I was shown:

warning C4244: '+=': conversion from 'double' to 'int', possible loss of data

Upvotes: 1

Related Questions