0

I was solving the leetcode problem called excel sheet column number using c++. Please see the link for the question -> Excel Sheet Column Number I came up with the following solution.

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

using namespace std;
#define val(c) (static_cast<int>(c) - 64)

int titleToNumber(string columnTitle) {
    int result = 0;
    int i = 0;
    int n = columnTitle.length() - 1;
    while(n >=0){
        cout<<"Current value = "<<val(columnTitle[i])*(pow(26,n))<<endl;
        result = result + val(columnTitle[i])*(pow(26,n));
        cout<<"Result = "<<result<<endl;
        n--;
        i++;
    }
    return result;
}
int main(){
string s  = "AA";
cout<<titleToNumber(s)<<endl;
return 0;
}

However, I am experiencing some unexpected error. I have initialized the value of result to 0 and when the program outputs the value of Current value it outputs it correctly which is 26, but when it adds and assigns the value to the variable result, the value somehow becomes 25. I checked the type of the value returned from the expression val(columnTitle[i])*(pow(26,n)) and it is of type double. I suspect that this is happening due to the type conversion from double to integer. However, the value of the expression is a whole number and does not have any floating part so there is no question of truncation. Also why the output value is 1 less than the expected?

Typecasting the expression value to integer doesn't help.

Can anyone please explain what is happening here and what I need to keep in mind in future for such conversions. Thank you.

Please note that when I run my solution on online c++ compiler such ashttps://www.onlinegdb.com/online_c++_compiler, the output is correct, but it is not correct when I run on my system. I am having g++.exe (MinGW.org GCC-6.3.0-1) 6.3.0 and I am also adding a snap of the output that I get. SNAP-output

Satish Kumar Singh
  • 117
  • 8
  • I get `Current value = 26 Result = 26` The next iteration of the loop is not what I expect but that's another question. – Martin York Nov 08 '22 at 07:54
  • The problem is that you're expecting `pow(26, x)` to be exactly an integer - there is no integral exponential in the standard library and you need to write your own (and look out for overflow). – molbdnilo Nov 08 '22 at 07:59
  • Yes on online compilers, it works fine but not on my pc. – Satish Kumar Singh Nov 08 '22 at 08:25
  • I suspect this is MinGW defect. I cannot reproduce this on any version of GCC, clang or MSVC I can get my hands on with any combination of optimisation/standard flags. – Revolver_Ocelot Nov 08 '22 at 08:25
  • @Revolver_Ocelot I am having g++.exe (MinGW.org GCC-6.3.0-1) 6.3.0 – Satish Kumar Singh Nov 08 '22 at 08:29
  • Make your life simple and use `std::round`. – Aziuth Nov 08 '22 at 08:34
  • Yah, I cannot get this behavior on pure GCC 6.3.0. GCC 6.3.0 is also ancient. I suggest switching either to MSVC or MinGW-w64 compiler. – Revolver_Ocelot Nov 08 '22 at 08:41
  • @SatishSingh None of this requires any floating point functions. Regardless of the compiler, your solution of using `pow` is broken. Are you going to change compilers if you later find out that `pow` gives you other issues? Do not use floating point functions to solve integer-based problems. – PaulMcKenzie Nov 08 '22 at 08:50
  • @PaulMcKenzie. Thnax. Changing the pow function solves it. – Satish Kumar Singh Nov 08 '22 at 09:09

0 Answers0