Replace the zeros in a NumPy integer array with nan

Question

I wrote a python script below:

import numpy as np

arr = np.arange(6).reshape(2, 3)
arr[arr==0]=['nan']
print arr

But I got this error:

Traceback (most recent call last):
  File "C:\Users\Desktop\test.py", line 4, in <module>
    arr[arr==0]=['nan']
ValueError: invalid literal for long() with base 10: 'nan'
[Finished in 0.2s with exit code 1]

How to replace zeros in a NumPy array with nan?

Answer 1

np.nan has type float: arrays containing it must also have this datatype (or the complex or object datatype) so you may need to cast arr before you try to assign this value.

The error arises because the string value 'nan' can't be converted to an integer type to match arr's type.

>>> arr = arr.astype('float')
>>> arr[arr == 0] = 'nan' # or use np.nan
>>> arr
array([[ nan,   1.,   2.],
       [  3.,   4.,   5.]])

Answer 2

An integer array can't hold a NaN value, so a new copy will have to be created anyway; so numpy.where may be used here to replace the values that satisfy the condition by NaN:

arr = np.arange(6).reshape(2, 3)
arr = np.where(arr==0, np.nan, arr)

# array([[nan,  1.,  2.],
#        [ 3.,  4.,  5.]])