import pandas as pd

# Read a CSV file and display its contents
df = pd.read_csv("data/sample_data.csv")
print(df.head())

Reading data from CSV files is a common task in data pipelines. You use the read_csv function from the pandas library to load the file into a DataFrame. This function automatically detects the delimiter (default is comma), but you can specify a different delimiter using the delimiter or sep parameter if your file uses something else, such as a tab or semicolon. File encoding is another important aspect; most CSV files use UTF-8 encoding, but you might encounter files with different encodings like ISO-8859-1. You can specify the encoding with the encoding parameter. If you try to read a file with the wrong encoding, you may see errors or garbled text. Error handling is crucial during extraction. The read_csv function provides options like error_bad_lines=False (deprecated in newer pandas versions) or on_bad_lines="skip" to skip problematic rows, and warn_bad_lines=True to display warnings. Always check the documentation for your pandas version to ensure you use the correct parameters.

import pandas as pd

# Read a JSON file with nested structures
df = pd.read_json("data/nested_data.json")

# If the JSON file contains deeply nested data, use json_normalize
if "records" in df.columns:
    from pandas import json_normalize
    nested_df = json_normalize(df["records"])
    print(nested_df.head())
else:
    print(df.head())