NumPy copysign() – Copy Sign from One Array to Another

NumPy copysign()

The numpy.copysign() function changes the sign of the elements in an input array to match the sign of another array or scalar.

Syntax

numpy.copysign(x1, x2, /, out=None, *, where=True, casting='same_kind', order='K', dtype=None, subok=True)

Parameters

Parameter	Type	Description
x1	array_like	The values whose sign needs to be changed.
x2	array_like	The sign of each element in x2 is copied to the corresponding element in x1. If x2 is a scalar, its sign is applied to all elements of x1.
out	ndarray, None, or tuple of ndarray and None, optional	Optional output array to store results. If not provided, a new array is created.
where	array_like, optional	Boolean mask specifying which elements to process. Unselected elements retain their original value.
casting	str, optional	Defines the casting behavior when applying the function.
order	str, optional	Memory layout order of the output array.
dtype	data-type, optional	Specifies the data type of the output array.
subok	bool, optional	Determines whether to preserve subclasses of ndarray in the output.

Return Value

Returns an array where the values of x1 have the same magnitude but the sign of x2. If both x1 and x2 are scalars, a scalar is returned.

---

Examples

1. Copying the Sign of a Scalar

Here, we change the sign of a single value based on another scalar.

import numpy as np

# Define values
x1 = -5.0  # Value to change
x2 = 3.0   # Sign donor

# Apply copysign
result = np.copysign(x1, x2)

# Print the result
print("Original value:", x1)
print("Sign donor:", x2)
print("Result after copysign:", result)

Output:

Original value: -5.0
Sign donor: 3.0
Result after copysign: 5.0

2. Applying copysign to an Array

Using numpy.copysign() to change the sign of multiple elements based on another array.

import numpy as np

# Define input arrays
x1 = np.array([1.0, -2.5, 3.8, -4.2])  # Values to modify
x2 = np.array([-1.0, 2.0, -3.0, 4.0])  # Sign donor

# Apply copysign
result = np.copysign(x1, x2)

# Print the results
print("Original values:", x1)
print("Sign donor values:", x2)
print("Result after copysign:", result)

Output:

Original values: [ 1.  -2.5  3.8 -4.2]
Sign donor values: [-1.  2. -3.  4.]
Result after copysign: [-1.   2.5 -3.8  4.2]

3. Using the out Parameter

Storing results in a predefined output array instead of creating a new array.

import numpy as np

# Define input values
x1 = np.array([-7.5, 8.2, -3.6])
x2 = np.array([2.0, -5.0, 4.0])

# Create an output array with the same shape
output_array = np.empty_like(x1)

# Apply copysign with the out parameter
np.copysign(x1, x2, out=output_array)

# Print results
print("Result after copysign:", output_array)

Output:

Result after copysign: [ 7.5 -8.2  3.6]

4. Using the where Parameter

Applying sign changes only to selected elements based on a condition.

import numpy as np

# Define input arrays
x1 = np.array([-5.0, 6.0, -7.0, 8.0])
x2 = np.array([1.0, -1.0, 1.0, -1.0])

# Define a mask to apply copysign conditionally
mask = np.array([True, False, True, False])

# Apply copysign using the where parameter
result = np.copysign(x1, x2, where=mask)

# Print the results
print("Result after copysign with mask:", result)

Output:

Result after copysign with mask: [5. 0. 7. 0.]

Only the elements where mask=True had their signs changed, while the rest remained unchanged.