Question

I am working with a multidimensional array but i get an exception, i have searched a lot but i find the same answer i’m using, the exception jumps when i try to allocate matriz[i] = new double[n]. I have tried both the commented and uncommented solutions with no luck.

void interpol(double *arr_x, double *arr_y, int n, double *results) {
    //double** matriz = new double*[n];
    double** matriz;
    matriz = (double**) malloc(n * sizeof(double*));
    for(int i = 0; i < n; i++){    
        //matriz[i] = new double[n+1];
        matriz[i] = (double*) malloc(n+1 * sizeof(double));
        for(int j = 0; j < n; j++) {
            matriz[i][j] = pow(arr_x[i],j);
        }
        matriz[i][n] = arr_y[i];
    }
    gaussiana(matriz, n, results);
}

— EDIT—

The function gaussiana is working fine, since i have tested outside this function. The exception is thrown in either:
//matriz[i] = new double[n];
matriz[i] = (double*) malloc(n * sizeof(double));

n is never more than 10.

The exception thrown is:

First-chance exception at 0x00071c4d in Interpolacion.exe: 0xC0000005:
Access violation reading location 0x00000000.
Unhandled exception at 0x774b15de in Interpolacion.exe: 0xC0000005: Access violation reading location 0x00000000.
The program ‘[8012] Interpolacion.exe: Native’ has exited with code -1073741819 (0xc0000005).

—-EDIT—-
I finally got it working, the issue was not in matriz, but with arr_x/arr_y, the external routine was sending the data wrong (oddly the error and the stacktrace always referred me to the new double[n] assignation)

Answer 1

If you want to use the std::vector route, you can use something like below (untested, shown just as a guide). Keep in mind that std::vector<std::vector<double> > is not compatible with double **, so your gaussiana function might need to be rewritten to accept the new type.:

// Include the header!
#include <vector>


// Be careful about the use of "using namespace std", I'm only using it here
// because it's a small example
using namespace std;

vector<vector<double> > matriz;

for (int i = 0; i < n; i++)
{
    // Create a new vector "v" with n+1 elements
    vector<double> v(n + 1);

    // fill this vector
    for (int j = 0; j < n; j++)
        v[j] = pow(arr_x[i], j);
    v[n] = arr_y[i];   

    // add it to the matrix
    matriz.push_back(v);
}