NumPy Input / Output (I/O)

In real-world data analysis, data rarely comes directly as NumPy arrays.

Most of the time, data is stored in files, such as text files, CSV files, or binary formats.

NumPy provides efficient input and output (I/O) functions to :

• Load data from files
• Save processed data
• Work with large datasets efficiently

Understanding NumPy I/O is essential for building real data pipelines.

Why NumPy I/O Is Important

NumPy I/O helps you :

• Read external data into arrays
• Save intermediate results
• Reuse processed data
• Improve performance when working with large files

Without I/O, analysis cannot move beyond small examples.

Saving NumPy Arrays to Files

Saving a Single Array Using save( )

import numpy as np
arr = np.array([10, 20, 30, 40])
np.save("data.npy", arr)

This saves the array in a binary format, which is :

• Fast
• Memory efficient
• Preserves data type and shape

Loading NumPy Arrays Using load( )

loaded_arr = np.load("data.npy")
print(loaded_arr)

The array is restored exactly as it was saved.

Saving Multiple Arrays Using savez()

a = np.array([1, 2, 3])
b = np.array([4, 5, 6])
np.savez("multiple_arrays.npz", arr1=a, arr2=b)

This is useful when working with :

• Multiple datasets
• Model inputs and outputs

Loading Multiple Arrays

data = np.load("multiple_arrays.npz")
print(data["arr1"])
print(data["arr2"])

Working with Text Files

Saving Data to a Text File

arr = np.array([100, 200, 300])
np.savetxt("values.txt", arr)

Text files are human-readable but :

• Larger in size
• Slower than binary files

Loading Data from a Text File

arr = np.loadtxt("values.txt")
print(arr)

Working with CSV Files

CSV files are very common in analytics.

Loading a CSV File

data = np.loadtxt("sales.csv", delimiter=",")
print(data)

Useful when :

• Working with exported reports
• Handling structured numerical data

Handling Missing Values in Files

data = np.loadtxt("data.csv", delimiter=",", skiprows=1)

This skips header rows when loading data.

Real-Life Example: Loading Sales Data

sales = np.loadtxt("monthly_sales.csv", delimiter=",")
total_sales = np.sum(sales)
print("Total Sales:", total_sales)

This is a common analytics workflow :

1. Load data
2. Analyze
3. Save results

Binary vs Text File Formats

Binary Files (.npy, .npz)

• Faster
• Smaller size
• Best for large datasets

Text Files (.txt, .csv)

• Human-readable
• Easy to share
• Slower for large data

Choosing the right format improves performance.

Common Beginner Mistakes

• Using text files for very large datasets
• Forgetting delimiters in CSV files
• Not handling headers correctly
• Overwriting important files

Understanding I/O prevents data loss and errors.

Why NumPy I/O Matters in Data Analytics

NumPy I/O is used in :

• Data preprocessing pipelines
• Machine learning workflows
• Intermediate result storage
• Performance optimization

Efficient data loading and saving saves time and resources.

Key Takeaways

• NumPy supports efficient file I/O
• Binary formats are faster and more reliable
• CSV and text files are useful for data sharing
• I/O is essential for real-world analytics

Conclusion

NumPy Input and Output functions make it easy to work with real-world data stored in files. They allow you to efficiently load data, save processed results, and reuse arrays without losing structure or performance. By choosing the right file format and using NumPy’s I/O tools correctly