numpy.nan_to_num() function replaces NaN (Not a Number) with a specified numerical value (default is 0), and optionally converts positive or negative infinity to finite numbers.

Example:

import numpy as np
a = np.array([1.0, np.nan, np.inf, -np.inf])
res = np.nan_to_num(a)
print(res)

[ 1.00000000e+000  0.00000000e+000  1.79769313e+308 -1.79769313e+308]

Explanation: By default, NaN is replaced with 0. Infinity values remain unchanged unless specified.

Syntax

numpy.nan_to_num(x, copy=True, nan=0.0, posinf=None, neginf=None)

Parameters:

• x: Input array (can be scalar, array-like).
• copy: If True (default), returns a copy. If False, modifies the array in-place.
• nan: Value to replace NaNs with (default is 0.0).
• posinf: Value to replace positive infinity with (optional).
• neginf: Value to replace negative infinity with (optional).

Returns: This method returns an array with the same shape as x, where NaN, +inf and -inf are replaced with finite numbers.

Examples

Example 1: Replace only NaNs with zero

import numpy as np
a = np.array([np.nan, 2, 3])
res = np.nan_to_num(a)
print(res)

[0. 2. 3.]

Explanation: NaN is replaced with 0, other values are unchanged.

Example 2: Replace NaN and infinite values

import numpy as np
a = np.array([np.nan, np.inf, -np.inf, 5])
res = np.nan_to_num(a, nan=0, posinf=1000, neginf=-1000)
print(res)

[    0.  1000. -1000.     5.]

Explanation: NaN → 0, +inf → 1000, -inf → -1000 and regular values like 5 remain unchanged.

Example 3: In-place modification with copy=False

import numpy as np
a = np.array([np.nan, np.inf])
np.nan_to_num(a, nan=-1, posinf=9999, copy=False)
print(a)

[-1.000e+00  9.999e+03]

Explanation: The original array 'a' is modified directly without making a copy.

Related articles: Numpy