In numpy, you can access the size and shape of the array with array.size and array.shape while elements can be accessed using array[i].
How does one achieve this with a C structure? One can do something like
struct{
int size;
int shape[2];
int *elements;
} int_array;
but then elements can be accessed as
int_array.elements[i].
not like numpy.
How can I have an array that stores size and shape, but whose elements can be accessed in the usual way using []?
3 *
sizeof(int).You can then access the array in the usual way.
Don't forget to subtract 3 * sizeof(int) from the pointer before you free it.
You could use a simple macro to cast your pointer to a 2D array (source):
#define INDEX(a) ((int(*)[a.shape[1]]) a.elements)
Which you can use as
#include <stdio.h>
#include <stdlib.h>
typedef struct{
int size;
int shape[2];
int *elements;
} int_array;
int main() {
int_array arr;
arr.shape[0] = 2;
arr.shape[1] = 3;
arr.elements = malloc(6*sizeof(*arr.elements));
INDEX(arr)[0][1] = 2;
printf("%d", INDEX(arr)[0][1]);
}
xtensor