NumPy ndarray.choose() Method

NumPy ndarray.choose()

The numpy.ndarray.choose() method constructs a new array by selecting elements from multiple choices based on an index array. Each element in the calling array serves as an index to pick a corresponding element from a set of given choices.

Syntax

</>

Copy

ndarray.choose(choices, out=None, mode='raise')

Parameters

Parameter	Type	Description
`choices`	sequence of arrays	A sequence of arrays from which elements are chosen. The shape of each choice array must match the shape of the calling array.
`out`	ndarray, optional	Alternative output array to store the result. It must match the expected shape.
`mode`	{‘raise’, ‘wrap’, ‘clip’}, optional	Determines behavior when an index is out of bounds: `'raise'` (default): Raises an error if an index is out of range. `'wrap'`: Wraps around using modular arithmetic. `'clip'`: Clips indices to the valid range.

Return Value

Returns an array constructed by choosing elements from choices according to the indices in the calling array.

Examples

1. Basic Usage of ndarray.choose()

In this example, we use choose() to select values from different arrays based on an index array.

</>

Copy

import numpy as np

# Define an index array
index_array = np.array([[0, 1], 
                         [1, 0]])

# Define the choice arrays
choice1 = np.array([[10, 20], 
                     [30, 40]])
choice2 = np.array([[100, 200], 
                     [300, 400]])

# Select values using choose()
result = index_array.choose([choice1, choice2])

print("Chosen elements:\n", result)

Output:

Chosen elements:
[[ 10 200]
 [300  40]]

The values are selected from choice1 if the index is 0 and from choice2 if the index is 1.

2. Using Out Parameter in ndarray.choose()

Here, we store the result in a predefined output array.

</>

Copy

import numpy as np

index_array = np.array([[1, 0], 
                         [0, 1]])

choice1 = np.array([[5, 10], 
                     [15, 20]])
choice2 = np.array([[50, 100], 
                     [150, 200]])

# Define an output array with the same shape as the expected result
output_array = np.empty_like(choice1)

# Store the result in the output array
index_array.choose([choice1, choice2], out=output_array)

print("Output array:\n", output_array)

Output:

Output array:
[[ 50  10]
 [ 15 200]]

The selected values are stored in output_array instead of creating a new one.

3. Handling Out-of-Bounds Indices with mode=’clip’

Using mode='clip' ensures that out-of-bounds indices are clipped to the valid range.

</>

Copy

import numpy as np

index_array = np.array([[0, 2],  # 2 is out of bounds
                         [1, -1]]) # -1 is also out of bounds

choice1 = np.array([[1, 2], 
                     [3, 4]])
choice2 = np.array([[10, 20], 
                     [30, 40]])
choice3 = np.array([[100, 200], 
                     [300, 400]])

# Applying choose() with mode='clip'
result = index_array.choose([choice1, choice2, choice3], mode='clip')

print("Result with mode='clip':\n", result)

Output:

Result with mode='clip':
[[  1 200]
 [ 30   1]]

Out-of-bounds indices are clipped to the nearest valid index. The index 2 is valid, but -1 is clipped to 0.

4. Using mode=’wrap’ to Wrap Indices

Using mode='wrap' wraps out-of-bounds indices using modular arithmetic.

</>

Copy

import numpy as np

index_array = np.array([[0, 3],  # 3 wraps around to 0
                         [1, -2]]) # -2 wraps around to 1

choice1 = np.array([[5, 10], 
                     [15, 20]])
choice2 = np.array([[50, 100], 
                     [150, 200]])

# Applying choose() with mode='wrap'
result = index_array.choose([choice1, choice2], mode='wrap')

print("Result with mode='wrap':\n", result)

Output:

Result with mode='wrap':
[[  5  100]
 [150 20]]

The index 3 wraps to 0, and -2 wraps to 1, ensuring valid selection.

TutorialKart

NumPy ndarray.choose() Method