CODEWITHSUNDEEP
pythonnumpyboolean
3sm1r
3sm1r

Reputation: 530

apply a boolean operator among all elements of an array

Is there a practical way to apply the same boolean operator (say or) to all elements of an array without using a for loop ?

I will clarify what I need with an example:

import numpy as np
a=np.array([[1,0,0],[1,0,0],[0,0,1]])
b=a[0] | a[1] | a[2]
print b

What is the synthetic way to apply the or boolean operator to all arrays of a matrix as I have done above?

Upvotes: 0

Views: 1825

Answers (4)

Dan Strohl
Dan Strohl

Reputation: 23

NOTE: I'm not a numby expert, so I am making an assumption below;

In your example, b is comparing the arrays with each other, so it is asking:

"Are there any items in a[0] OR any items in a1 OR any items in a2" is that your goal? in which case, you could use builtin any()

for example (changed numby to a simple list of lists):

a=[[1,0,0],[1,0,0],[0,0,1]]
b=any(a)
print b

b will be True

if however, you want to know if any element in it is true, so, for example, you want to know if a[0][0] OR a0 | a0 | a[1][0] | ...

you could use the builtin map command, so something like:

a=[[1,0,0],[1,0,0],[0,0,1]]
b=any(map(any, a)
print b

b will still be True

Note: below is based on looking at the NumPy docs, not actual experience.

For NumPy, you could also use the NumPy any() option something like

a=np.array([[1,0,0],[1,0,0],[0,0,1]])
b=a.any()
print b

or, if your doing all numbers anyway, you could sum the array and see if it != 0

Upvotes: 0

user2357112
user2357112

Reputation: 281381

The usual way to do this would be to apply numpy.any along an axis:

numpy.any(a, axis=0)

That said, there is also a way to do this through the operator more directly. NumPy ufuncs have a reduce method that can be used to apply them along an axis of an array, or across all elements of an array. Using numpy.logical_or.reduce, we can express this as

numpy.logical_or.reduce(a, axis=0)

This doesn't come up much, because most ufuncs you'd want to call reduce on already have equivalent helper functions defined. add has sum, multiply has prod, logical_and has all, logical_or has any, maximum has amax, and minimum has amin.

Upvotes: 3

Derek Eden
Derek Eden

Reputation: 4628

try either:

np.any(arr, axis=0)

or

np.apply_along_axis(any, 0, arr)

or if you want to use pandas for some reason,

df.any(axis=0)

Upvotes: 1

Seba Arriagada
Seba Arriagada

Reputation: 380

You can use reduce function for that:

from functools import reduce

a = np.array([[1,0,0],[1,0,0],[0,0,1]])
reduce(np.bitwise_or, a)

Upvotes: 0

Related Questions