How to check for NaNs in a numpy ndarray which contains strings

Question

Normally I can check for NaNs with numpy.isnan() :

import numpy as np

arr1 = np.array([[1, 2], [3, 4], [np.nan, 5]])
print(type(arr1))
print(np.isnan(arr1))

<class 'numpy.ndarray'>
[[False False]
 [False False]
 [ True False]]

But how can I achieve the same for an array which contains strings?

arr2 = np.array([[1, 2], [3, 4], [np.nan, 5], ['high', 'low']])
print(type(arr2))
print(np.isnan(arr2))

<class 'numpy.ndarray'>

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-19-4a4a4bc72076> in <module>
      1 array = np.array([[1, 2], [3, 4], [np.nan, 5], ['high', 'low']])
      2 print(type(array))
----> 3 print(np.isnan(array))

TypeError: ufunc 'isnan' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''

What does the 2nd array look luke.? you printed it? If an array of strings, test for the 'nan' string — hpaulj, Apr 01 '20 at 11:51

score 0 · Answer 1 · answered Apr 01 '20 at 14:18

As suggested by the commentors, the inclusion of strings in the array has led to all elements to be silently coerced to strings, including the np.nan values which are now 'nan', and can be found with arr2==nan:

print(arr1=='nan')

[[False False]
 [False False]
 [ True False]
 [False False]]

How to check for NaNs in a numpy ndarray which contains strings

1 Answers1