Den 27.01.2011 22:03, skrev Dewald Pieterse: > Is numpy.append so slow? or is the culprit numpy.where? Please observe that appending to a Python list is amortized O(1), whereas appending to a numpy array is O(N**2). Sturla