How do you count the amount of zero rows of a numpy array?
array = np.asarray([[1,1],[0,0],[1,1],[0,0],[0,0]])
-> has three rows with all zero's, hence should give 3
Took me sometime to figure out this one and also couldn't find an answer on SO
Add a comment
|
3 Answers
You could also leverage the "truthiness" of non-zero values in an array.
np.sum(~array.any(1))
i.e., sum the rows where none of the values in said row are truthy (and hence are all zero)
Comments
You can use np.all if you're sure that the rows will have all zeros.
# number of rows which has non-zero elements
In [33]: sum(np.all(arr, axis=1))
Out[33]: 2
# invert the mask to get number of rows which has all zero as its elements
In [34]: sum(~np.all(arr, axis=1))
Out[34]: 3
1 Comment
miradulo
This is incorrect for cases where a row contains at least one zero but not all zeros.
import numpy as np
array = np.array([[1,1],[0,0],[1,1],[0,0],[0,0]])
non_zero_rows = np.count_nonzero((array != 0).sum(1)) # gives 2
zero_rows = len(array) - non_zero_rows # gives 3
Any better solutions?