Get the first column of a matrix represented by a vector of vectors

As I mentioned in the comments, it’s not practical to represent matrices using vector-of-vector for a few reasons:

1. It is fiddly to set up;
2. It is difficult to change;
3. Cache locality is bad.

Here is a very simple class I have created that will hold a 2D matrix in a single vector. This is pretty much how software like MATLAB does it… albeit a huge simplification.

template <class T>
class SimpleMatrix
{
public:
    SimpleMatrix( int rows, int cols, const T& initVal = T() );

    // Size and structure
    int NumRows() const                       { return m_rows; }
    int NumColumns() const                    { return m_cols; }
    int NumElements() const                   { return m_data.size(); }

    // Direct vector access and indexing
    operator const vector<T>& () const        { return m_data; }
    int Index( int row, int col ) const       { return row * m_cols + col; }

    // Get a single value
          T & Value( int row, int col )       { return m_data[Index(row,col)]; }
    const T & Value( int row, int col ) const { return m_data[Index(row,col)]; }
          T & operator[]( size_t idx )        { return m_data[idx]; }
    const T & operator[]( size_t idx ) const  { return m_data[idx]; }

    // Simple row or column slices
    vector<T> Row( int row, int colBegin = 0, int colEnd = -1 ) const;
    vector<T> Column( int col, int rowBegin = 0, int rowEnd = -1 ) const;

private:
    vector<T> StridedSlice( int start, int length, int stride ) const;

    int m_rows;
    int m_cols;

    vector<T> m_data;
};

This class is basically sugar-coating around a single function — StridedSlice. The implementation of that is:

template <class T>
vector<T> SimpleMatrix<T>::StridedSlice( int start, int length, int stride ) const
{
    vector<T> result;
    result.reserve( length );
    const T *pos = &m_data[start];
    for( int i = 0; i < length; i++ ) {
        result.push_back(*pos);
        pos += stride;
    }
    return result;
}

And the rest is pretty straight-forward:

template <class T>
SimpleMatrix<T>::SimpleMatrix( int rows, int cols, const T& initVal )
    : m_data( rows * cols, initVal )
    , m_rows( rows )
    , m_cols( cols )
{    
}

template <class T>
vector<T> SimpleMatrix<T>::Row( int row, int colBegin, int colEnd ) const
{
    if( colEnd < 0 ) colEnd = m_cols-1;
    if( colBegin <= colEnd )
        return StridedSlice( Index(row,colBegin), colEnd-colBegin+1, 1 );
    else
        return StridedSlice( Index(row,colBegin), colBegin-colEnd+1, -1 );
}
    
template <class T>
vector<T> SimpleMatrix<T>::Column( int col, int rowBegin, int rowEnd ) const
{
    if( rowEnd < 0 ) rowEnd = m_rows-1;
    if( rowBegin <= rowEnd )
        return StridedSlice( Index(rowBegin,col), rowEnd-rowBegin+1, m_cols );
    else
        return StridedSlice( Index(rowBegin,col), rowBegin-rowEnd+1, -m_cols );
}

Note that the Row and Column functions are set up in such a way that you can easily request an entire row or column, but are a little more powerful because you can slice a range by passing one or two more parameters. And yes, you can return the row/column in reverse by making your start value larger than your end value.

There is no bounds-checking built into these functions, but you can easily add that.

You could also add something to return an area slice as another SimpleMatrix<T>.

Have fun.