Replace Negative Value with Zero in NumPy Array

If an array has negative numbers, you change them to 0, while keeping all non-negative numbers the same. For example: Suppose you have this array: A=[−3,  5,  −7,  2,  0,  −1], after replacing negatives with zero: A′=[0,  5,  0,  2,  0,  0].

Using boolean mask (in-place)

This is the most efficient method if you want to update the original array directly. It uses NumPy’s boolean indexing.

Example: In this example, we create a boolean mask for all negative elements and assign them the value 0 directly inside the same array.

import numpy as np

a = np.array([1, 2, -3, 4, -5, -6])
print("Input:", a)

a[a < 0] = 0
print("Result:", a)

Input: [ 1  2 -3  4 -5 -6]
Result: [1 2 0 4 0 0]

Explanation:

• a < 0: produces a boolean mask identifying negative positions.
• a[a < 0] = 0: selects those positions and assigns 0 in-place.
• The original array a is modified; no new array is created.

Using np.maximum 

This method uses NumPy’s universal function (ufunc) to compare each element with 0.

Example: Here we apply np.maximum to take the larger of each element and 0, ensuring negatives are replaced.

import numpy as np

a = np.array([1, 2, -3, 4, -5, -6])
res = np.maximum(a, 0)
print("Input:", a)
print("Result:", res)

Input: [ 1  2 -3  4 -5 -6]
Result: [1 2 0 4 0 0]

Explanation:

• np.maximum(a, 0) compares each element of a with 0.
• For each position it returns the larger of the two values effectively replacing negatives with 0.
• res is a new array (original a unchanged).

Using np.clip

np.clip clamps all values to lie within a given range. Setting the lower bound to 0 forces negatives up to 0.

Example: In this example, we clip values of the array to the range [0, upper], so anything below 0 becomes 0.

import numpy as np

a = np.array([1, 2, -3, 4, -5, -6])
res = np.clip(a, 0, max(a.max(), 0))
print("Input:", a)
print("Result:", res)

Input: [ 1  2 -3  4 -5 -6]
Result: [1 2 0 4 0 0]

Explanation:

• a.max() finds the largest element; max(a.max(), 0) ensures the upper bound ≥ 0.
• np.clip(a, 0, upper) forces every value below 0 to become 0, leaving others unchanged.
• The result is a new array.

Using np.where

np.where lets you select between two values based on a condition.

Example: This code creates a new array: if an element is negative, it becomes 0; otherwise, it keeps its original value.

import numpy as np

a = np.array([1, 2, -3, 4, -5, -6])
res = np.where(a < 0, 0, a)
print("Input:", a)
print("Result:", res)

Input: [ 1  2 -3  4 -5 -6]
Result: [1 2 0 4 0 0]

Explanation:

• a < 0 produces a boolean mask.
• np.where(mask, 0, a) chooses 0 where mask is True, otherwise the original element.
• Returns a new array; dtype follows broadcasting rules.

Using np.vectorize

This method wraps a Python function to apply elementwise. It is slower but sometimes useful for readability.

Example: In this program, a lambda function replaces negative elements with 0 and keeps others unchanged.

import numpy as np

a = np.array([1, 2, -3, 4, -5, -6])
vfunc = np.vectorize(lambda x: 0 if x < 0 else x)
res = vfunc(a)
print("Input:", a)
print("Result:", res)

Input: [ 1  2 -3  4 -5 -6]
Result: [1 2 0 4 0 0]

Explanation:

• np.vectorize applies the lambda to each element.
• For negatives, lambda returns 0. Otherwise, it returns the value unchanged.
• The result is correct but slower than ufunc-based methods like maximum or clip.