Hi I’m trying to perform a search operation in an array that contains multiple 2d arrays comparing it’s itens to a specific array. I managed to do it using a for loop iterating trough the itens inside the big array but I have to perform this search 10^6 times and the length of this for loop can grow up to 2^100 so it is becoming very time consuming. I’m wondering if there is a way to make this search faster using a np.where or np.isin() function.
this is an example of the slower working method
import numpy as np frequencies = {} b = np.array ([[0, 0, 0], [0, 0, 0], [1, 1, 1]]) #template a = np.array([[[1, 1, 1], [0, 0, 0], [0, 0, 0]], [[0, 0, 0], [1, 1, 1], [0, 0, 0]],[[0, 0, 0], [0, 0, 0], [1, 1, 1]],[[0, 0, 1], [0, 0, 1], [0, 0, 1]],[[0, 1, 0], [0, 1, 0], [0, 1, 0]],[[1, 0, 0], [1, 0, 0], [1, 0, 0]]]) #I need to know if b is inside a and the index where it its located for I in range (a): if np.all(b==a[I]): frequencies [I] = frequencies [I] + 1
and I would like to make something like this. I need to store the indexes where b is found inside a in the index c of the dictionary frequencies
import numpy as np frequencies = {} b = np.array ([[0, 0, 0], [0, 0, 0], [1, 1, 1]]) #template a = np.array([[[1, 1, 1], [0, 0, 0], [0, 0, 0]], [[0, 0, 0], [1, 1, 1], [0, 0, 0]],[[0, 0, 0], [0, 0, 0], [1, 1, 1]],[[0, 0, 1], [0, 0, 1], [0, 0, 1]],[[0, 1, 0], [0, 1, 0], [0, 1, 0]],[[1, 0, 0], [1, 0, 0], [1, 0, 0]]]) c = np.where(np.all(b==a)) frequencies [c] = frequencies [c] + 1
Advertisement
Answer
You can use NumPy.all
with a two-axis then use NumPy.argwhere
for finding the index like below:
b = np.array ([[0, 0, 0], [0, 0, 0], [1, 1, 1]]) a = np.array([[[1, 1, 1], [0, 0, 0], [0, 0, 0]], [[0, 0, 0], [1, 1, 1], [0, 0, 0]],[[0, 0, 0], [0, 0, 0], [1, 1, 1]],[[0, 0, 1], [0, 0, 1], [0, 0, 1]],[[0, 1, 0], [0, 1, 0], [0, 1, 0]],[[1, 0, 0], [1, 0, 0], [1, 0, 0]]]) np.all(b == a, axis=(-1)) # array([[False, True, False], # [ True, False, False], # [ True, True, True], # [False, False, False], # [False, False, False], # [False, False, False]]) np.all(b == a, axis=(-1,1)) # array([False, False, True, False, False, False]) indexs = np.argwhere(np.all(b == a, axis=(-1, 1)))
Output:
>>> indexs array([[2]])