Read Matrix from file, give it to main - c++

Question

I have a fairly simple program. It reads a matrix from a file that looks like this
4 4
a b c d
a b c d
a b c d
a b c d
and outputs it in the console. The program is working as intended:
#include <iostream>
#include <fstream>
using namespace std;
int main() {
int n, m;
ifstream myfile;
myfile.open ("matrix.txt");
myfile >> n >> m;
char mat[n][m];
for (int i = 0; i < n; i++) {
for (int j = 0; j < m; j++) {
myfile >> mat[i][j];
}
}
for (int i = 0; i < n; i++) {
for (int j = 0; j < m; j++) {
cout << mat[i][j];
}
cout << endl;
}
return 0;
}
Now this program is going to grow a lot, so I want to start using headerfiles and declare the functions external so that my main looks more like
int main() {
readMat();
printMat();
}
My Problem is, I don't really know how to do this right. Where should I declare the matrix so that both functions can "see" it? I can't declare it globally as char mat[n][m] because I only know n and m inside readMat().
Or is my whole design flawed and there is much better way to do this?
I would be grateful for every little tip.
I have not worked with multiple files in c++ yet.

Answer 1

You cannot use an array like that, when not just multiple translation units, but multiple functions are involved.
Any function needs to know the exact data types it receives as parameters, and this must be declared at compile time.
Here, you don't know the type of the array until runtime. array[10] and array[15] are different types. Don't be fooled by the fact their name might be the same. It's not the name that matters, but the type, like int [10] or int [11].
You have two basic options:
Use templates.
Convert your array to a std::vector<std::vector<int>>. Then you can declare the array (now a vector) globally, and/or pass it as a parameter to your multiple functions.

Answer 2

Matrix.h
#ifndef MATRIX_H
#define MATRIX_H
namespace mynamespace {
template <class T>
class Matrix {
private:
int m_rows;
int m_columns;
T** m_data;
public:
Matrix(); // Default Constructor - Null or Empty Matrix
Matrix( const int rows, const int columns ); // Empty Matrix With Defined Size
Matrix( const int rows, const int columns, const T** data ); // Defined Matrix
T** getData() const;
}; // Matrix
#include "Matrix.inl"
} // mynamespace
#endif // MATRIX_H
Matrix.inl
// Matrix Constructor and Function Implementations Here
Matrix.cpp
#include "Matrix.h"
Main.cpp
#include <string>
#include <iostream>
#include <fstream>
#include "Matrix.h"
using namespace mynamespace;
void readMatrix( std::fstream& fin, const std::string& strFilename, Matrix& matrixIn );
void printMatrix( const Matrix& matrix );
int main() {
std::fstream fileIn; // Prefer to not use "using namespace std;"
std::string strFilename( "Matrix.txt" );
Matrix<type> mat(); // Default Empty Matrix To Be Filled In
readMatrix( fileIn, strFilename, mat );
printMatrix( mat );
return 0;
} // main
void readMatrix( std::fstream& fin, const std::string& strFilename, Matrix& mat ) {
// Do Work Here To Populate Matrix Object
} // readMatrix
void printMatrix( Matrix& mat ) {
// Do Work Here To Print Matrix Being Passed In
} // printMatrix
With this pseudo code, the design here allows any type of matrix to be constructed. The matrix is an object of a template type. The implementation of generating a matrix by reading in data from a file, and the implementation of displaying or printing that matrix are independent of the class itself. This allows for the matrix class to be as generic as possible without having to rely on any other libraries. The functions that do the work are stand alone functions that accept a matrix object. A modification may be needed in your text or binary file to work with this construct.
matrix.txt
type
4 4
a b c d
a b c d
a b c d
a b c d
The only problem with this, is it should already be obvious: you need to define what type of data this matrix will hold when instantiating it. Yet you do not know the type until after you read the file in. So this code does have a bug in it that would need to be address. However, the overall design method used here is to only show how to keep similar code objects contained with in their own class object while separating independent code blocks that does work on them. Now other class objects may have a file reading and print methods, but normally Vector (math & not containers) and Matrix classes usually don't.