Indexing and Slicing with NumPy

Indexing and Slicing with NumPy

Basic Indexing and Slicing Concepts

Indexing and slicing are your tools for extracting specific data from NumPy arrays. Think of indexing as pointing to exact locations, while slicing grabs entire sections in one go. Unlike Python lists that create copies, NumPy slicing creates views that share memory with the original array.

This memory-sharing behavior makes NumPy efficient but requires careful attention to avoid unintended modifications. The numpy.org provides extensive examples of advanced indexing patterns when you need more detailed data selection.

One-Dimensional Array Operations

One-dimensional arrays use the same syntax as Python lists. Slicing returns a view of the data, not a copy, so changes to the view will affect the original array.

indexing_1d.py

import numpy as np# Create 1D arraya = np.linspace(0, 7, 8)print("Original array:", a)  # Output: Original array: [0. 1. 2. 3. 4. 5. 6. 7.]# Single element indexingprint("3rd element:", a[3])  # Output: 3rd element: 3.0print("Last element:", a[-1])  # Output: Last element: 7.0# Basic slicingprint("Slice [2:6]:", a[2:6])  # Output: Slice [2:6]: [2. 3. 4. 5.]print("Slice [3:-2]:", a[3:-2])  # Output: Slice [3:-2]: [3. 4. 5.]# Modification through slicing (changes original array)a[:3] = 0print("After a[:3] = 0:", a)  # Output: After a[:3] = 0: [0. 0. 0. 3. 4. 5. 6. 7.]

view_vs_copy.py

import numpy as np# Original arrayoriginal = np.array([1, 2, 3, 4, 5])print("Original:", original)  # Output: Original: [1 2 3 4 5]# Create view through slicingview = original[1:4]print("View:", view)  # Output: View: [2 3 4]# Modify view (affects original)view[0] = 999print("After modifying view:")print("Original:", original)  # Output: Original: [  1 999   3   4   5]print("View:", view)  # Output: View: [999   3   4]# Create explicit copycopy_array = original[1:4].copy()copy_array[0] = 777print("After modifying copy:")print("Original:", original)  # Output: Original: [  1 999   3   4   5]print("Copy:", copy_array)  # Output: Copy: [777   3   4]

indexing_2d.py

import numpy as np# Create 2D arraya = np.array([[1., 2., 3.],            [4., 5., 6.],            [7., 8., 9.],            [10., 11., 12.]])print("2D Array:")print(a)# Output:# [[ 1.  2.  3.]#  [ 4.  5.  6.]#  [ 7.  8.  9.]#  [10. 11. 12.]]# Specific element indexingprint("a[2,1]:", a[2,1])  # Output: a[2,1]: 8.0# Row slicingprint("a[1,:]:", a[1,:])  # Output: a[1,:]: [4. 5. 6.]# Column slicingprint("a[:,2]:", a[:,2])  # Output: a[:,2]: [ 3.  6.  9. 12.]

advanced_slicing_2d.py

import numpy as np# 2D array for demonstrationa = np.array([[1., 2., 3.],            [4., 5., 6.],            [7., 8., 9.],            [10., 11., 12.]])# Subarray slicingprint("a[1:3, 0:2]:")print(a[1:3, 0:2])# Output:# [[4. 5.]#  [7. 8.]]# Slicing with stepprint("a[::2, :]:")print(a[::2, :])# Output:# [[ 1.  2.  3.]#  [ 7.  8.  9.]]# Column slicing with stepprint("a[:, ::2]:")print(a[:, ::2])# Output:# [[ 1.  3.]#  [ 4.  6.]#  [ 7.  9.]#  [10. 12.]]

Operation	Description	Example Result
`a[2,1]`	Element at row $2$ , column $1$	Single value: `8.0`
`a[1,:]`	Entire row $1$	One-dimensional array: `[4. 5. 6.]`
`a[:,2]`	Entire column $2$	One-dimensional array: `[3. 6. 9. 12.]`
`a[1:3, 0:2]`	Subarray rows $1\text{-}2$ , columns $0\text{-}1$	Two-dimensional array: `[[4. 5.], [7. 8.]]`
`a[::2, :]`	Every second row	Two-dimensional array with rows $0, 2$
`a[:, ::2]`	Every second column	Two-dimensional array with columns $0, 2$

integer_array_indexing.py

import numpy as np# Array for demonstrationa = np.array([0, 1, 2, 3, 4])print("Original array:", a)  # Output: Original array: [0 1 2 3 4]# Indexing with list of indicesi = [1, 3, 2, 1, 4]print("Index array:", i)  # Output: Index array: [1, 3, 2, 1, 4]print("a[i]:", a[i])  # Output: a[i]: [1 3 2 1 4]# Indexing and reshaping with 2D index arrayi = np.array([[1, 2], [3, 4]])print("2D index array:")print(i)# Output:# [[1 2]#  [3 4]]print("a[i]:")print(a[i])# Output:# [[1 2]#  [3 4]]

boolean_mask_indexing.py

import numpy as np# Array for demonstrationa = np.linspace(0, 5, 6)print("Array:", a)  # Output: Array: [0. 1. 2. 3. 4. 5.]# Create boolean maskmask = np.array([True, False, True, False, True, False])print("Boolean mask:", mask)  # Output: Boolean mask: [ True False  True False  True False]print("a[mask]:", a[mask])  # Output: a[mask]: [0. 2. 4.]# Boolean mask from conditioncondition_mask = a % 2 == 0print("Condition mask (a % 2 == 0):", condition_mask)  # Output: Condition mask (a % 2 == 0): [ True False  True False  True False]print("a[condition_mask]:", a[condition_mask])  # Output: a[condition_mask]: [0. 2. 4.]# Boolean mask with complex conditionscomplex_mask = (a > 1) & (a < 4)print("Complex mask (a > 1) & (a < 4):", complex_mask)  # Output: Complex mask (a > 1) & (a < 4): [False False  True  True False False]print("a[complex_mask]:", a[complex_mask])  # Output: a[complex_mask]: [2. 3.]

combined_indexing.py

import numpy as np# 2D array for demonstrationa = np.arange(24).reshape(4, 6)print("2D Array:")print(a)# Output:# [[ 0  1  2  3  4  5]#  [ 6  7  8  9 10 11]#  [12 13 14 15 16 17]#  [18 19 20 21 22 23]]# Combination of slicing and indexingprint("a[1:3, [0, 2, 5]]:")print(a[1:3, [0, 2, 5]])# Output:# [[ 6  8 11]#  [12 14 17]]# Boolean indexing on rows, slicing on columnsrow_mask = np.array([True, False, True, False])print("a[row_mask, 2:5]:")print(a[row_mask, 2:5])# Output:# [[ 2  3  4]#  [14 15 16]]

step_slicing.py

import numpy as np# Array for demonstrationa = np.arange(20)print("Array:", a)  # Output: Array: [ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19]# Slicing with stepprint("a[::2] (every second element):", a[::2])  # Output: a[::2] (every second element): [ 0  2  4  6  8 10 12 14 16 18]print("a[1::3] (start index 1, every third):", a[1::3])  # Output: a[1::3] (start index 1, every third): [ 1  4  7 10 13 16 19]print("a[::-1] (reverse array):", a[::-1])  # Output: a[::-1] (reverse array): [19 18 17 16 15 14 13 12 11 10  9  8  7  6  5  4  3  2  1  0]# Step on 2D arrayb = np.arange(12).reshape(3, 4)print("2D Array:")print(b)# Output:# [[ 0  1  2  3]#  [ 4  5  6  7]#  [ 8  9 10 11]]print("b[::2, ::2] (every second row and column):")print(b[::2, ::2])# Output:# [[ 0  2]#  [ 8 10]]

ellipsis_slicing.py

import numpy as np# 3D array for demonstrationa = np.arange(24).reshape(2, 3, 4)print("3D Array shape:", a.shape)  # Output: 3D Array shape: (2, 3, 4)# Using ellipsisprint("a[0, ...] (same as a[0, :, :]):")print(a[0, ...])# Output:# [[ 0  1  2  3]#  [ 4  5  6  7]#  [ 8  9 10 11]]print("a[..., 2] (same as a[:, :, 2]):")print(a[..., 2])# Output:# [[ 2  6 10]#  [14 18 22]]print("a[1, ..., ::2] (same as a[1, :, ::2]):")print(a[1, ..., ::2])# Output:# [[12 14]#  [16 18]#  [20 22]]

boolean_operations.py

import numpy as np# Array for demonstrationa = np.linspace(0, 5, 6)print("Array:", a)  # Output: Array: [0. 1. 2. 3. 4. 5.]# Manual boolean maskmask = np.array([True, False, True, False, True, False], dtype=bool)print("Manual mask:", mask)  # Output: Manual mask: [ True False  True False  True False]print("a[mask]:", a[mask])  # Output: a[mask]: [0. 2. 4.]# Boolean mask from comparisongreater_than_2 = a > 2print("a > 2:", greater_than_2)  # Output: a > 2: [False False False  True  True  True]print("a[a > 2]:", a[a > 2])  # Output: a[a > 2]: [3. 4. 5.]# Boolean mask with compound conditionseven_and_greater_than_1 = (a % 2 == 0) & (a > 1)print("(a % 2 == 0) & (a > 1):", even_and_greater_than_1)  # Output: (a % 2 == 0) & (a > 1): [False False  True False  True False]print("a[even_and_greater_than_1]:", a[even_and_greater_than_1])  # Output: a[even_and_greater_than_1]: [2. 4.]

boolean_2d.py

import numpy as np# 2D array for demonstrationa = np.arange(12).reshape(3, 4)print("2D Array:")print(a)# Output:# [[ 0  1  2  3]#  [ 4  5  6  7]#  [ 8  9 10 11]]# Boolean mask for specific elementsmask = a > 5print("Mask a > 5:")print(mask)# Output:# [[False False False False]#  [False False  True  True]#  [ True  True  True  True]]print("a[mask]:", a[mask])  # Output: a[mask]: [ 6  7  8  9 10 11]# Boolean indexing with assignmenta[a < 5] = 0print("After a[a < 5] = 0:")print(a)# Output:# [[ 0  0  0  0]#  [ 0  0  6  7]#  [ 8  9 10 11]]

fancy_indexing_1d.py

import numpy as np# Array for demonstrationa = np.array([10, 20, 30, 40, 50, 60])print("Array:", a)  # Output: Array: [10 20 30 40 50 60]# Fancy indexing with list of indicesindices = [0, 2, 4, 1]print("Indices:", indices)  # Output: Indices: [0, 2, 4, 1]print("a[indices]:", a[indices])  # Output: a[indices]: [10 30 50 20]# Fancy indexing with NumPy arraynp_indices = np.array([5, 1, 3, 1, 0])print("NumPy indices:", np_indices)  # Output: NumPy indices: [5 1 3 1 0]print("a[np_indices]:", a[np_indices])  # Output: a[np_indices]: [60 20 40 20 10]# Fancy indexing with 2D index arrayindices_2d = np.array([[0, 1], [2, 3]])print("2D Indices:")print(indices_2d)# Output:# [[0 1]#  [2 3]]print("a[indices_2d]:")print(a[indices_2d])# Output:# [[10 20]#  [30 40]]

fancy_indexing_2d.py

import numpy as np# 2D array for demonstrationa = np.arange(24).reshape(4, 6)print("2D Array:")print(a)# Output:# [[ 0  1  2  3  4  5]#  [ 6  7  8  9 10 11]#  [12 13 14 15 16 17]#  [18 19 20 21 22 23]]# Fancy indexing for specific rowsrow_indices = [0, 2, 3]print("a[row_indices, :]:")print(a[row_indices, :])# Output:# [[ 0  1  2  3  4  5]#  [12 13 14 15 16 17]#  [18 19 20 21 22 23]]# Fancy indexing for specific elementsrow_idx = [0, 1, 2, 3]col_idx = [1, 2, 3, 4]print("a[row_idx, col_idx]:", a[row_idx, col_idx])  # Output: a[row_idx, col_idx]: [ 1  8 15 22]# Combination of fancy indexing with slicingprint("a[[0, 2], 1:4]:")print(a[[0, 2], 1:4])# Output:# [[ 1  2  3]#  [13 14 15]]

data_filtering.py

import numpy as np# Temperature sensor data simulationtemperatures = np.array([22.5, 25.1, 19.8, 30.2, 18.5, 27.3, 31.1, 24.8])print("Temperature data:", temperatures)  # Output: Temperature data: [22.5 25.1 19.8 30.2 18.5 27.3 31.1 24.8]# Filter normal temperatures (20-28 degrees)normal_temp_mask = (temperatures >= 20) & (temperatures <= 28)normal_temps = temperatures[normal_temp_mask]print("Normal temperatures:", normal_temps)  # Output: Normal temperatures: [22.5 25.1 27.3 24.8]# Filter extreme temperaturesextreme_temp_mask = (temperatures < 20) | (temperatures > 30)extreme_temps = temperatures[extreme_temp_mask]print("Extreme temperatures:", extreme_temps)  # Output: Extreme temperatures: [19.8 30.2 18.5 31.1]# Replace extreme values with averagemean_temp = temperatures[normal_temp_mask].mean()temperatures_cleaned = temperatures.copy()temperatures_cleaned[extreme_temp_mask] = mean_tempprint("Data after cleaning:", temperatures_cleaned)  # Output: Data after cleaning: [22.5   25.1   24.925 24.925 24.925 27.3   24.925 24.8  ]

data_sampling.py

import numpy as np# Simulation datasetdata = np.arange(100).reshape(10, 10)print("Dataset shape:", data.shape)  # Output: Dataset shape: (10, 10)# Random row samplingnp.random.seed(42)sample_indices = np.random.choice(10, size=5, replace=False)sample_indices.sort()print("Sample indices:", sample_indices)  # Output: Sample indices: [0 1 5 7 8]sampled_data = data[sample_indices, :]print("Sampled data shape:", sampled_data.shape)  # Output: Sampled data shape: (5, 10)print("First 3 columns of sampled data:")print(sampled_data[:, :3])# Output:# [[ 0  1  2]#  [10 11 12]#  [50 51 52]#  [70 71 72]#  [80 81 82]]# Column reorganizationcolumn_order = [9, 0, 5, 2, 7, 1, 8, 3, 6, 4]reorganized_data = data[:, column_order]print("Reorganized first row:", reorganized_data[0, :])  # Output: Reorganized first row: [9 0 5 2 7 1 8 3 6 4]