> ## Documentation Index
> Fetch the complete documentation index at: https://python4ai.codewithsiva.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Working with files

> Read, process, and write data files

## Files in AI projects

When working with Python for AI, you'll constantly work with data files. Your data might come as:

* **CSV files** - Spreadsheet data from Excel or databases
* **JSON files** - API responses and configuration data
* **XML files** - Structured data from various systems
* **Text files** - Raw text for processing
* **Parquet files** - Efficient data storage format

The good news? Python has excellent libraries for all of these.

## Common libraries for files

Each file type has specialized libraries:

**CSV files:**

* `pandas` - Best for data analysis (recommended)
* `csv` module - Built-in, for simple operations

**JSON files:**

* `json` module - Built-in, handles all JSON operations
* `pandas` - Can read/write JSON with DataFrames

**Other formats:**

* `xml.etree` - Built-in XML parsing
* `openpyxl` - Excel files (.xlsx)
* `PyPDF2` - PDF files

## Working with our sales data

Let's work with our CSV file and convert it to different formats. First, install pandas:

```bash theme={null}
pip install pandas
```

<Note>
  If you get an error, try `pip3 install pandas` or install it through VS Code's terminal.
</Note>

Update your `analyzer.py`:

```python theme={null}
import pandas as pd
import json
import os

# Read the CSV file
df = pd.read_csv('data/sales.csv')
print("CSV Data:")
print(df)
print(f"\nShape: {df.shape[0]} rows, {df.shape[1]} columns")

# Quick operation: calculate total for each row
df['total'] = df['quantity'] * df['price']
print("\nWith totals:")
print(df)

# Create output directory
os.makedirs('output', exist_ok=True)

# Save as different formats
# 1. JSON format (good for web APIs)
df.to_json('output/sales_data.json', orient='records', indent=2)

# 2. Excel format (good for sharing)
df.to_excel('output/sales_data.xlsx', index=False)

# 3. Updated CSV (with our new total column)
df.to_csv('output/sales_with_totals.csv', index=False)

print("\nFiles saved:")
print("- output/sales_data.json")
print("- output/sales_data.xlsx") 
print("- output/sales_with_totals.csv")
```

## File format comparison

Different formats have different uses:

```python theme={null}
# JSON - Great for APIs and web applications
{
  "date": "2024-01-01",
  "product": "Laptop",
  "quantity": 2,
  "price": 999.99
}

# CSV - Simple, universal, good for data analysis
date,product,quantity,price
2024-01-01,Laptop,2,999.99

# Excel - Feature-rich, good for business users
# (Binary format with formatting, formulas, etc.)
```

## Loading different file types

Here's how to load various formats:

```python theme={null}
# CSV
df = pd.read_csv('data/file.csv')

# JSON
df = pd.read_json('data/file.json')
# or for simple JSON:
with open('data/config.json', 'r') as f:
    data = json.load(f)

# Excel
df = pd.read_excel('data/file.xlsx')

# Text files
with open('data/file.txt', 'r') as f:
    text = f.read()
```

## Learn more

To dive deeper into file handling:

* [Pandas documentation](https://pandas.pydata.org/docs/) - Comprehensive data handling
* [Python JSON module](https://docs.python.org/3/library/json.html) - Official JSON docs
* [Real Python file I/O](https://realpython.com/working-with-files-in-python/) - Detailed tutorial
* [CSV module docs](https://docs.python.org/3/library/csv.html) - Built-in CSV handling

<Card title="Organizing code" icon="arrow-right" href="/practical-python/organizing-code">
  Split your code into reusable functions
</Card>