Declaring an array inside a function

[jad123xp]

Hello, I am making a program that takes in an unspecified amount of integers inside an array and adds up the absolute values of the numbers. Look at the code to understand more.

#include <iostream>
#include <cmath>
#include <windows.h>

using namespace std;

int getAbsSum(int nums[]){
    int sum = 0;
    for(int i = 0; i <= (sizeof(nums) / sizeof(int)); i++){
        if(nums[i] < 0){
            nums[i] = 0 - nums[i];
            sum += nums[i];
        } else {
            sum += nums[i];
        }
    }
    return sum;
}

int main()
{
    cout << getAbsSum([1, -4, -5, 6, 1]);

    return 0;
}

The error is

|22|error: expected identifier before numeric constant
|22|error: expected ']' before ',' token
|22|error: expected '{' before ',' token
|22|error: expected ')' before ']' token

[kigar64551]

You can't create an array "inline" like that in C/C++. Instead, try the following:

int main()
{
    int input[] = { 1, -4, -5, 6, 1 };
    cout << getAbsSum(input);
}

...or maybe have a look at:
https://cplusplus.com/reference/initializer_list/initializer_list/

[deleted account xyzzy]

Using std::vector as the dynamic array, initialized with initializer list syntax in the function call, range-based for loops and std::abs:

#include <iostream>
#include <cmath>
#include <vector>

int getAbsSum(const std::vector<int>& nums)
{
   int sum {};

   for (const auto& i : nums)
   {
      sum += std::abs(i);
   }
   return sum;
}

int main()
{
   std::cout << getAbsSum({ 1, -4, -5, 6, 1 }) << '\n';
}

17

[deleted account xyzzy]

It is possible to pass a regular array of "unknown size" into a function using a template, but the array has to be declared first:

#include <iostream>
#include <cmath>

template <size_t N>
int getAbsSum(int (&nums)[N])
{
   int sum {};

   for (size_t i {}; i < N; ++i)
   {
      sum += std::abs(nums[i]);
   }
   return sum;
}

int main()
{
   int nums[] { 1, -4, -5, 6, 1 };

   std::cout << getAbsSum(nums) << '\n';
}

I'd stick with using C++ containers when possible instead of using regular arrays, they make "programming life" much easier.

[anup30]

#include <iostream>
#include <numeric> //for std::accumulate(),
using namespace std;

int main(){
	 auto abs_sum = [](int a, int b){ //lambda function	
	 	return b<0? a-b : a+b;
	};  
	 int ar[]={1,-2,3,4};
	 int size= sizeof(ar)/sizeof(ar[0]);
	 cout<< endl << accumulate(ar, ar+size, 0, abs_sum);
}

10

[mbozzi]

#include <iostream>

int getAbsSum(int nums[], int size){
    int sum = 0;

    for (int i = 0; i < size; ++i)
    {
        if (nums[i] >= 0)
        {
            sum += nums[i]; 
        }
        else
        {
            sum -= nums[i];
        }
    }

    return sum;
}

int main()
{
    int xs[] = {1, -4, -5, 6, 1};
    std::cout << getAbsSum(xs, (sizeof xs / sizeof *xs));

    return 0;
}

[mbozzi]

@anup30
Try std::transform_reduce or std::inner_product instead:

std::transform_reduce(std::begin(xs), std::end(xs), 0, std::plus{}, 
  [](auto x){ return x < 0? -x: x; });

[jad123xp]

Thanks everyone.
It was solved.
i am kind of a beginner here.
so thank you very much

[anup30]

@mbozzi, can you write the full the program?(because i dont understand why you are using transform_reduce)
and my program works fine. do i even need to change it? while it has no bugs or malfunctions?

[mbozzi]

@mbozzi, can you write the full the program?

#include <iterator>
#include <numeric>
#include <iostream>
#include <functional>

int main() 
{
  int const xs[] { 1, -4, -5, 6, 1 };
  int const sum = std::transform_reduce(std::begin(xs), std::end(xs), 0, std::plus{}, 
    [](auto x){ return x < 0? -x: x; });
  std::cout << sum << '\n';
}

Conceptually, each element is "transformed" by computing its absolute value. Then the list of absolute values is "reduced" into a single value by adding them together.

My program works fine. Do i even need to change it? while it has no bugs or malfunctions?

Of course not, yours is fine. But given the problem description
"Sum the absolute values of an array",
transform_reduce or inner_product is a better fit.

[anup30]

did you run the program?

[deleted account xyzzy]

anup30 wrote:

did you run the program?

Did you?

Runs for me using Visual Studio 2022 set to language standard C++20.

A hint, std::transform_reduce requires C++17 or better to work, std::inner_product is a bit more relaxed.

https://en.cppreference.com/w/cpp/algorithm/transform_reduce
https://en.cppreference.com/w/cpp/algorithm/inner_product

[lastchance]

1 2 3 4 5 6 7 8 9 10 11 12 13

#include <iostream> #include <valarray> using namespace std; int getAbsSum( const valarray<int> & V ) { return abs( V ).sum(); } int main() { cout << getAbsSum( { 1, -4, -5, 6, 1 } ); }

17