How is numpy's fancy indexing implemented

NumPy's fancy indexing allows you to select elements from an array using arrays of indices or boolean masks. It provides a powerful way to access and manipulate elements in arrays. Fancy indexing is implemented efficiently in NumPy, typically using C or Cython under the hood for optimized performance.

Let's go through five examples of fancy indexing and explain each step in detail:

Example 1: Basic Fancy Indexing

import numpy as np

# Creating an array
arr = np.array([1, 2, 3, 4, 5])

# Fancy indexing
indices = [0, 2, 4]
result = arr[indices]

print(result)

Explanation:

1. We import the NumPy library.
2. We create an array arr containing [1, 2, 3, 4, 5].
3. We define indices as [0, 2, 4], indicating the indices we want to select.
4. Using fancy indexing arr[indices], we select elements at indices 0, 2, and 4.
5. The selected elements [1, 3, 5] are returned as result.

Output: [1 3 5]

Example 2: Fancy Indexing with 2D Array

import numpy as np

# Creating a 2D array
arr = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])

# Fancy indexing
indices = [0, 2]
result = arr[:, indices]

print(result)

Explanation:

1. We create a 2D array arr.
2. arr is:

   [[1 2 3]
    [4 5 6]
    [7 8 9]]

1. We define indices as [0, 2], indicating the columns we want to select.
2. Using fancy indexing arr[:, indices], we select all rows (:) and columns at indices 0 and 2.
3. The selected columns [[1 3] [4 6] [7 9]] are returned as result.

Output:

[[1 3]
 [4 6]
 [7 9]]

Example 3: Fancy Indexing with Boolean Mask

import numpy as np

# Creating an array
arr = np.array([1, 2, 3, 4, 5])

# Boolean mask
mask = np.array([True, False, True, False, True])
result = arr[mask]

print(result)

Explanation:

1. We create an array arr containing [1, 2, 3, 4, 5].
2. We define a boolean mask mask where True indicates the elements to be selected.
3. Using fancy indexing arr[mask], we select elements where the mask is True.
4. The selected elements [1, 3, 5] are returned as result.

Output: [1 3 5]

Example 4: Combining Fancy Indexing and Slicing

import numpy as np

# Creating a 2D array
arr = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])

# Fancy indexing with slicing
indices = [0, 2]
result = arr[:2, indices]

print(result)

Explanation:

1. We create a 2D array arr.
2. We define indices as [0, 2].
3. Using fancy indexing and slicing arr[:2, indices], we select the first two rows and columns at indices 0 and 2.
4. The selected elements [[1 3] [4 6]] are returned as result.

Output:

[[1 3]
 [4 6]]

Example 5: Fancy Indexing with Repeating Indices

import numpy as np

# Creating an array
arr = np.array([1, 2, 3, 4, 5])

# Fancy indexing with repeating indices
indices = [0, 0, 2, 2, 4, 4]
result = arr[indices]

print(result)

Explanation:

1. We create an array arr containing [1, 2, 3, 4, 5].
2. We define indices with repeating indices.
3. Using fancy indexing arr[indices], NumPy repeats the elements at specified indices.
4. The selected elements [1, 1, 3, 3, 5, 5] are returned as result.

Output: [1 1 3 3 5 5]

These examples demonstrate the versatility and power of fancy indexing in NumPy, allowing you to select elements from arrays in various ways to suit your needs.