numpy.nansum() function computes the sum of array elements over a given axis, treating NaN (Not a Number) values as zero. This is useful when you want to ignore missing or undefined values in your computation.

For Example:

import numpy as np
a = np.array([1.0, 2.0, np.nan, 4.0])
res = np.nansum(a)
print(res)

7.0

Explanation: np.nansum() ignores the NaN and returns the sum of 1.0 + 2.0 + 4.0 = 7.0.

Syntax

numpy.nansum(a, axis=None, dtype=None, keepdims=<no value>)

Parameters:

• a: Input array to compute the sum over.
• axis: Axis or axes along which the sum is computed; default is all elements.
• dtype: Desired data type of the returned array.
• keepdims: If True, retains reduced dimensions with size one.

Returns: This function returns the sum of array elements (ignoring NaNs), optionally over the specified axis.

Examples of numpy.nansum()

Example 1: Applying along rows using axis=1

import numpy as np
a = np.array([[1, 2], [np.nan, 3]])
res = np.nansum(a, axis=1)
print(res)

[3. 3.]

Explanation:

• Row 0: 1 + 2 = 3
• Row 1: NaN is ignored, so 3 is returned

Example 2: Using keepdims=True

import numpy as np
a = np.array([[1, 2], [np.nan, 3]])
res = np.nansum(a, axis=1, keepdims=True)
print(res)

[[3.]
 [3.]]

Explanation: The result keeps the original shape along the reduced axis (1-column instead of reducing to 1D array).

Example 3: Specifying data type with dtype

import numpy as np
a = np.array([1, 2, 3], dtype=np.int32)
res = np.nansum(a, dtype=np.float64)
print(res)

6.0

Explanation: Computation is done in float64 and the result is the sum 1 + 2 + 3 = 6.