I\'ve been skimming through the Armadillo documentation and examples, but it seems there is no real efficient way to subsample (or resample) a large vector or matrix, such t
