I'm starting to love Eigen, and quite new to this, but suppose I find eigenvectors of a matrix type MatrixXd

using SelfAdjointEigenSolver<MatrixXd> eigensolver ..... and I get eigenvectors via eigensolver.eigenvectors();

then how can I turn that into standard c++ matrix form? i.e. how can I push back each eigenvector into vector<double> ?

You can map both data structures as a linear 1D array:

vector<double> vec(n*n);
VectorXd::Map(vec.data(), n*n) = VectorXd::Map(eigensolver.eigenvectors().data(), n*n);

or use memcpy...