根据它们在另一个二维数组中的存在从二维数组中删除子数组（返回一个新的二维数组）

Question

I have two 2d array, A and B, like the following.

A=array([[1, 1, 1],
       [2, 2, 2],
       [3, 3, 3],
       [4, 4, 4],
       [5, 5, 5],
       [6, 6, 6],
       [7, 7, 7],
       [8, 8, 8],
       [9, 9, 9]])
B=array([[1, 1, 1],
       [3, 3, 3],
       [8, 8, 8]])

I want to remove subarrays of A if they exist in B. Then return a new 2d array C like the following:

C=array([[2, 2, 2],
       [4, 4, 4],
       [5, 5, 5],
       [6, 6, 6],
       [7, 7, 7],
       [9, 9, 9]])

Currently I have tried the np.isin function but the result is no longer a 2d array.

mask = np.isin(A, B, invert=True)
A[mask]
>>array([2, 2, 2, 4, 4, 4, 5, 5, 5, 6, 6, 6, 7, 7, 7, 9, 9, 9])

Answer 1

You need to aggregate the booleans on axis 1:

C = A[~np.isin(A,B).all(1)]

output:

array([[2, 2, 2],
       [4, 4, 4],
       [5, 5, 5],
       [6, 6, 6],
       [7, 7, 7],
       [9, 9, 9]])

Answer 2

Or in1d :

C = A[~np.in1d(A, B).all(axis=1)]

And now:

print(C)

Output:

[[2 2 2]
 [4 4 4]
 [5 5 5]
 [6 6 6]
 [7 7 7]
 [9 9 9]]