vector transpose question

The code below transfers the two dimensional array in_arr and produces the output in out_arr. I then intialize a vector to look like in_arr. I'd like to do the same thing with the vector. What woudl the code look like in the transponse function for the vector?
Thanks in advance.

Code:

  
# include <vector> 
# include <algorithm>
using namespace std; // this is bad but only here for test purposes

template <typename T>
void  transpose(std::vector<T>& input,
	         std::vector<T>& output)  // these probably need to be const correct
{
  for ( size_t idx (0); idx < input.size(); ++idx )
  {
    //output [ idx + 1 ] = input [ idx + 1 ];
  }
}


int main()
{
  int in_arr  [ 3 ] [ 3 ];
  int out_arr [ 3 ] [ 3 ];

  // initialize arr;
  int temp ( 4 );
  for ( int idx ( 0 ); idx < 3; ++idx )
    for ( int jdx ( 0 ); jdx < 3; ++jdx )
      in_arr [ idx ] [ jdx ] = temp * jdx;
  // dump in
  for ( int idx ( 0 ); idx < 3; ++idx )
    for ( int jdx ( 0 ); jdx < 3; ++jdx )
      cout << "  " << in_arr [ idx ][ jdx ] << endl;

  cout << endl;

  // transpose
  for ( int idx ( 0 ); idx < 3; ++idx )
    for ( int jdx ( 0 ); jdx < 3; ++jdx )
      out_arr [ jdx ][ idx ] = in_arr [ idx ] [ jdx ];

  // dump
  for ( int idx ( 0 ); idx < 3; ++idx )
    for ( int jdx ( 0 ); jdx < 3; ++jdx )
      cout << "  " << out_arr [ idx ][ jdx ] << endl;

  // now do the same with a vector
  vector<double> in;
  in.push_back ( 0 );
  in.push_back ( 4 );
  in.push_back ( 8 );
  in.push_back ( 0 );
  in.push_back ( 4 );
  in.push_back ( 8 ); 
  in.push_back ( 0 );
  in.push_back ( 4 );
  in.push_back ( 8 );

  vector<double> out;

  transpose<double> ( in, out );
  std::copy ( out.begin(), out.end(), 
              std::ostream_iterator<double> ( std::cout, "\n" ));

}