Numpy xor-shrink array

Question

How to xor all elements of a boolean numpy array using vectorized methods: i.e. a_1 xor a_2 xor ... xor a_n

?

+3

Yariv 13 Feb '13 at 9:21

2 answers

Probably most efficient to use sum

:

np.sum(arr) % 2

+2

ecatmur 13 Feb '13 at 9:49

seberg · Accepted Answer · 2013-02-13T11:55:11+0000

I would rather use xor ufunc, I think that bitwise_xor

(or logical_xor

):

np.bitwise_xor.reduce(a)

or

np.logical_xor.reduce(a)

One benefit is that you don't get bogus things to float.