45

Their are many links to go the other way round but I am unable to find to get a std::vector from a Eigen::Matrix or Eigen::VectorXd in my specific case.

Manish
  • 668
  • 1
  • 6
  • 13
  • Since I had a hard time finding the "many links to go the other way round", consider looking at https://stackoverflow.com/a/44613025/10220019 for the other way around. I hope this saves some people some time – C. Binair Feb 13 '23 at 13:11

3 Answers3

60
vector<int> vec(mat.data(), mat.data() + mat.rows() * mat.cols());
John Zwinck
  • 239,568
  • 38
  • 324
  • 436
  • 8
    `mat.rows() * mat.cols()` can be simplified to `mat.size()`, however, be aware that this solution only work for a plain `Matrix<>` object, while using a `Map<>` as in my answer works for sub-matrices too. – ggael Sep 29 '14 at 10:58
  • 3
    If someone needs to copy a row of a matrix into a vector: `vector vec(arr.cols()); Map(&vec[0], 1, mat.cols()) = mat.row(0);` – zardosht Jun 14 '20 at 17:57
58

You cannot typecast, but you can easily copy the data:

VectorXd v1;
v1 = ...;
vector<double> v2;
v2.resize(v1.size());
VectorXd::Map(&v2[0], v1.size()) = v1;
ggael
  • 28,425
  • 2
  • 65
  • 71
2

You can do this from and to Eigen vector : 

    //init a first vector
    std::vector<float> v1;
    v1.push_back(0.5);
    v1.push_back(1.5);
    v1.push_back(2.5);
    v1.push_back(3.5);

    //from v1 to an eignen vector
    float* ptr_data = &v1[0];
    Eigen::VectorXf v2 = Eigen::Map<Eigen::VectorXf, Eigen::Unaligned>(v1.data(), v1.size());

    //from the eigen vector to the std vector
    std::vector<float> v3(&v2[0], v2.data()+v2.cols()*v2.rows());


    //to check
    for(int i = 0; i < v1.size() ; i++){
        std::cout << std::to_string(v1[i]) << " | " << std::to_string(v2[i]) << " | " << std::to_string(v3[i]) << std::endl;
    }
Bastienm
  • 363
  • 2
  • 16