Matrix as function input - c

Question

I want to construct a function that takes a matrix as input and then calculates the means of all rows (or columns). I tried to do it like this, but it gives me a bunch of errors.
#include <stdio.h>
#include <stdlib.h>
int sum_rows[1000], sum_columns[1000], mean_row[1000], mean_column[1000], N, A[100][100], i, j;
void values(int A[N][N]) {
for (i=0; i<N; ++i) {
for (j=0; j<N; ++j) {
A[i][j] = (rand()%61+20);
}
}
}
void mean_r (int A[N][N]) {
for (i=0; i<N; ++i) {
sum_row[i] = 0;
for (j=0; j<N; ++j) {
sum_row[i] += A[i][j];
}
mean_row[i] = sum_row[i]/N;
}
}
void mean_c (int A[N][N]) {
for (j=0; j<N; ++j) {
sum_column[j] = 0;
for (i=0; i<N; ++i) {
sum_column[j] += A[i][j];
}
mean_column[j] = sum_column[j]/N;
}
}
int main()
{
int N;
printf("Enter size of matrix: ");
scanf("%d", &N);
int A[N][N];
values(A[N][N]);
mean_r(A[N][N]);
mean_c(A[N][N]);
for (i=0; i<N; ++i) {
for (j=0; j<N; ++j) {
printf("A[%d][%d] = %d\n", i, j, A[i][j]);
}
}
for (i=0; i<N; ++i) {
printf("Mean row #%d = %.2f\n", i, mean_row[i]);
}
for (j=0; j<N; ++j) {
printf("Promedio Columna #%d = %.2f\n", j, mean_column[j]);
}
return 0;
}
Can you explain to me how to pass a matrix into a function? I have to do these operations with a function defined beforehand, and just call it inside the main function.

Answer 1

It seems that your matrix is dynamic(the number of of rows and columns are not fixed). In C language, you are not allowed to declare the array of array with the dynamic rows and columns because the compiler does not know how to allocate the how much memory for you. You will have to allocate the memory at run-time. In my example, the matrix is stored in the memory as first row, the second row and so on. In math, the access to the element of the matrix is by row index and column index. "row index and column index" is translated to the memory address of the element using the starting address of the memory allocated + row number * column number + the column index based on how the matrix is stored in my example. Note: both row number and column number start with zero. The first element is denoted by p where p+0*columncount+0) is p. The next element is denoted by p+0*columncount+1. Following this, the element of the second row and second column(2,2) is denoted by p+1*column count+1. The following example uses single pointer to represent the dynamic matrix(Other technique does exist, eg. the pointer array)
In my example, 3(three rows) by 4(four columns) matrix is created and initialized. The ProcessMatrix function display the sum of each row.
If possible, I recommend you use C++ that allows you to allocate the memory with the size specified by the value of one integer, which makes things more easier. If you have access to other math libraries(many available), you should use them instead of trying to reinvent wheel.
// c_matrix.cpp : This file contains the 'main' function. Program execution begins and ends there.
//
#include <stdio.h>
int GetElementValue(int* m, int rowIndex, int colIndex, int columnCount)
{
return *(m + (rowIndex * columnCount + colIndex));
}
void ProcessMatric(int* myMatrix, int rowcount, int colcount)
{
// calculate the sum of each row
int sum = 0;
for (int i = 0; i < rowcount; i++)
{
sum = 0;
for (int j = 0; j < colcount; j++)
{
sum = sum + GetElementValue(myMatrix, i, j, colcount);
}
printf("Sum of row %d=%d\r\n", i+1, sum);
}
}
int main()
{
int row = 3;
int col = 4;
int* matrix;
//create the memory for the matrix 3 by 4.
matrix = malloc(sizeof(int) * row * col);
// in momery the matrix is stored as (row1,col1), (row1, col2), (row1,col3)...(row4,col1),(row4,col2),(row4, col3)
// initialize the matrix as 0,1,2,3...11
for (int i = 0; i < row * col; ++i)
{
*(matrix + i) = i;
}
ProcessMatric(matrix, row, col);
free(matrix);
return 0;
}

Answer 2

In your main function you declare an NxN matrix A
int A[N][N];
As Yan pointed out, you are unable to declare the array in such a way because the compiler doesn't know the value of N. The value of N won't be determined until after the program has been compiled and is executing (i.e. during runtime). Beyond that, though, it seems like what you want to do is pass the 2d array. What you need to pass is A itself (e.g. values(A);, however you pass
values(A[N][N]);
mean_r(A[N][N]);
mean_c(A[N][N]);
What you've done is index both dimensions, so, normally what would happen is you would just be passing an int (whatever is in the Nth row and Nth column). However, indices are in the range [0, N-1] so A[N] is out-of-range. They key concept is that arrays (and other data structures) are passed by reference. When you declare A, contiguous space will be allocated to hold N * N * sizeof(int) and the virtual address for the starting location of that space is stored in A. As a result, your functions know what memory addresses to store stuff in so that you can access it with A[i][j] without having to pass a giant slab of memory.