Eigen cannot create vector from matrix mean directly

Question

I am new to Eigen, and I would like to normalize a matrix in rowwise, so my code goes like this:

int buf[9];
for (int i{0}; i < 9; ++i) {
    buf[i] = i;
}
m = Map<MatrixXi>(buf, 3,3);
MatrixXi mean = m.colwise().mean();
VectorXi m2 = Map<VectorXi>(mean.data(), mean.cols());
m.rowwise() -=  m2;

This will not work, since m2 is interpreted as vertical, what is the cause of this ?

By the way, I just found that I could not avoid creating a mean matrix, which I thinks I could:

// this works
MatrixXi mean = m.colwise().mean();
VectorXi m2 = Map<VectorXi>(mean.data(), mean.cols());
// this cannot pass the compilation check
VectorXi m2 = Map<VectorXi>(m.colwise().mean().data(), m.cols());

What maybe the cause of this then ?

Answer 1

Your question is not very clear but I guess you're looking for .transpose(). Also no need to remap the result of .mean():

Map<MatrixXi> m(buf, 3,3);
VectorXi mean = m.colwise().mean();
m.rowwise() -=  mean.transpose();

or directly use a row vector:

RowVectorXi mean = m.colwise().mean();
m.rowwise() -=  mean;

or even the one-liner:

m.rowwise() -= m.colwise().mean().eval();