NumPy Sorting with Fancy Indexing
Fancy indexing in NumPy is a powerful technique that allows you to access and reorder array elements using arrays of indices.
When combined with sorting, fancy indexing helps you:
- Reorder data based on sorted positions
- Extract custom sorted views
- Build advanced data manipulation logic
What is Fancy Indexing?
Fancy indexing means using arrays of indices instead of single values or slices.
Example:
arr[[3, 1, 0]]
This selects elements in custom order.
Why Use Fancy Indexing for Sorting?
- Custom sorting control
- Works with complex datasets
-
Combines well with
argsort() - Useful in data science and ML
- Allows reordering based on conditions
1. Basic Sorting with Fancy Indexing
We first use argsort() to get sorted indices.
import numpy as np
arr = np.array([50, 10, 40, 20])
sorted_indices = np.argsort(arr)
print(sorted_indices)
print(arr[sorted_indices])
Output
[1 3 2 0]
[10 20 40 50]
2. How It Works
Original array: [50, 10, 40, 20]
Sorted indices: [ 1, 3, 2, 0]
Sorted result: [10, 20, 40, 50]
3. Descending Order Sorting
import numpy as np
arr = np.array([50, 10, 40, 20])
sorted_indices = np.argsort(arr)[::-1]
print(arr[sorted_indices])
Output
[50 40 20 10]
4. Fancy Indexing with 2D Arrays
import numpy as np
arr = np.array([
[3, 1, 2],
[9, 5, 6]
])
indices = np.argsort(arr, axis=1)
print(indices)
print(np.take_along_axis(arr, indices, axis=1))
Output
[[1 2 0]
[1 2 0]]
[[1 2 3]
[5 6 9]]
5. Sorting Rows Using Fancy Indexing
import numpy as np
arr = np.array([30, 10, 20, 40])
sorted_arr = arr[np.argsort(arr)]
print(sorted_arr)
Output
[10 20 30 40]
6. Reordering Based on Custom Index
import numpy as np
arr = np.array([100, 200, 300, 400])
order = [3, 1, 0, 2]
print(arr[order])
Output
[400 200 100 300]
7. Sorting Strings with Fancy Indexing
import numpy as np
arr = np.array(["banana", "apple", "cherry"])
indices = np.argsort(arr)
print(arr[indices])
Output
['apple' 'banana' 'cherry']
8. Sorting Rows by a Column Value
import numpy as np
data = np.array([
[3, 100],
[1, 50],
[2, 150]
])
sorted_data = data[data[:, 1].argsort()]
print(sorted_data)
Output
[[ 1 50]
[ 3 100]
[ 2 150]]
9. Real-World Example: Student Marks Sorting
import numpy as np
students = np.array(["A", "B", "C"])
marks = np.array([85, 70, 95])
sorted_students = students[np.argsort(marks)]
print(sorted_students)
Output
['B' 'A' 'C']
10. Real-World Example: Sales Ranking
import numpy as np
sales = np.array([500, 200, 900, 300])
ranked_sales = sales[np.argsort(sales)[::-1]]
print(ranked_sales)
Output
[900 500 300 200]
Key Concepts Summary
| Concept | Description |
|---|---|
| Fancy Indexing | Using index arrays |
| argsort() | Returns sorted indices |
| take_along_axis | Advanced sorting tool |
| [::-1] | Reverse order |
Advantages of Fancy Index Sorting
- Powerful custom sorting
- Works with multi-dimensional arrays
- Easy data rearrangement
- Useful in ML pipelines
- High flexibility for datasets
Summary
NumPy fancy indexing allows advanced sorting by using index arrays instead of direct values. Combined with argsort(), it enables powerful data reordering and structured dataset manipulation.
This functionality is part of NumPy and is widely used in applications built with Python.
Conclusion
Fancy indexing is an advanced NumPy technique that makes sorting and data manipulation highly flexible. It is especially useful in data science, machine learning, and structured data analysis.
0 Comments