Boolean Indexing

Creating a boolean array

A boolean array can be created manually by using dtype=bool when creating the array. Values other than 0, None, False or empty strings are considered True.

import numpy as np

bool_arr = np.array([1, 0.5, 0, None, 'a', '', True, False], dtype=bool)
print(bool_arr)
# output: [ True  True False False  True False  True False]

Alternatively, numpy automatically creates a boolean array when comparisons are made between arrays and scalars or between arrays of the same shape.

arr_1 = np.random.randn(3, 3)
arr_2 = np.random.randn(3, 3)

bool_arr = arr_1 < 0.5    
print(bool_arr.dtype)
# output: bool

bool_arr = arr_1 < arr_2
print(bool_arr.dtype)
# output: bool