Masked values in numpy digitize

Question

I want that numpy digitize ignores some values in my array. To achieve this I replaced the unwanted values by NaN and masked the NaN values:

import numpy as np
A = np.ma.array(A, mask=np.isnan(A))

Nonetheless np.digitize throws the masked values out as -1. Is there an alternative way so that np.digitize ignores the masked values (or NaN)?

score 2 · Accepted Answer · answered Nov 19 '19 at 16:10

I hope its not meant to be a performance optimization otherwise you can just mask after the digitize function:

import numpy as np

A = np.arange(10,dtype=np.float)
A[0] = np.nan
A[-1] = np.nan

bins = np.array([1,2,7])

res = np.digitize(A,bins)

# here np.nan is assigned to the highes bin 
# using numpy '1.17.2'
print(res)

# sp you mask you array after the execution of 
# np.digitize
print(res[~np.isnan(A)])

>>> [3 1 2 2 2 2 2 3 3 3]
>>> [1 2 2 2 2 2 3 3]

Masked values in numpy digitize

1 Answers1