2
votes

Python's round() seems to always round up when faced with x.5 numbers:

print round(1.5),round(2.5),round(3.5),round(4.5)
>>> 2.0 3.0 4.0 5.0

But numpy.round() seems to be inconsistent:

import numpy as np
print np.round(1.5),np.round(2.5),np.round(3.5),np.round(4.5)
>>> 2.0 2.0 4.0 4.0

This can introduce errors in certain cases. Is this a bug or am I missing something?

1
It's rounding to the nearest even integer for faster binary calculation. - Mikael
It's worth noting that this behaviour changed between Python 2 and Python 3; Python 3 behaves the same as NumPy here. - Mark Dickinson
@Mikael: I don't think speed has anything to do with it. - Mark Dickinson

1 Answers

5
votes

numpy rounds to the nearest even value:

https://docs.scipy.org/doc/numpy/reference/generated/numpy.around.html#numpy.around

For values exactly halfway between rounded decimal values, NumPy rounds to the nearest even value. Thus 1.5 and 2.5 round to 2.0, -0.5 and 0.5 round to 0.0, etc.