NumPy Record Arrays
Record arrays in NumPy are an extension of structured arrays that make data access easier and more readable.
Instead of using bracket notation like:
data["name"]
You can use dot notation:
data.name
This makes code cleaner and more intuitive.
What are Record Arrays?
Record arrays are structured arrays with attribute-style access to fields.
They allow you to treat structured data like objects.
Why Use Record Arrays?
- Cleaner syntax
- Easier field access
- More readable code
- Works like objects
- Useful in data analysis
1. Creating a Record Array
We use np.rec.array() or view().
import numpy as np
data = np.rec.array([
("Alice", 25, 50000),
("Bob", 30, 60000),
("Charlie", 28, 55000)
], dtype=[("name", "U10"), ("age", "i4"), ("salary", "i4")])
print(data)
2. Accessing Fields (Dot Notation)
print(data.name)
print(data.age)
Output
['Alice' 'Bob' 'Charlie']
[25 30 28]
3. Accessing Single Record
print(data[0])
Output
('Alice', 25, 50000)
4. Creating Record Array from Structured Array
import numpy as np
structured = np.array([
("John", 40, 80000),
("Emma", 29, 65000)
], dtype=[("name", "U10"), ("age", "i4"), ("salary", "i4")])
record = structured.view(np.recarray)
print(record.name)
5. Updating Values in Record Arrays
data.salary[0] = 52000
print(data)
6. Filtering Record Arrays
filtered = data[data.age > 27]
print(filtered)
7. Sorting Record Arrays
sorted_data = np.sort(data, order="age")
print(sorted_data)
8. Modifying String Fields
data.name = np.char.upper(data.name)
print(data.name)
9. Real-World Example: Employee Records
import numpy as np
employees = np.rec.array([
("John", 40, 80000),
("Emma", 29, 65000),
("Mike", 33, 72000)
], dtype=[("name", "U10"), ("age", "i4"), ("salary", "i4")])
print(employees.salary)
10. Real-World Example: Student Data
students = np.rec.array([
("A", 90),
("B", 75),
("C", 88)
], dtype=[("name", "U10"), ("marks", "i4")])
print(students.marks)
Record Array vs Structured Array
| Feature | Structured Array | Record Array |
|---|---|---|
| Access style | data["name"] | data.name |
| Readability | Medium | High |
| Performance | Fast | Similar |
| Use case | Data processing | Clean coding |
Advantages of Record Arrays
- Easy dot notation access
- Cleaner code
- Better readability
- Good for object-like data
- Useful in data science workflows
Summary
NumPy record arrays provide a more user-friendly way to access structured data using dot notation. They simplify code and make structured data handling more intuitive.
This feature is part of NumPy and is widely used in applications built with Python.
Conclusion
Record arrays are a powerful enhancement over structured arrays, making data access simpler and more readable. They are especially useful in data science and analytics projects.
0 Comments