I/O Operations

Code

import pandas as pd
from sqlalchemy import create_engine

df = pd.DataFrame({"name": ["A", "B"], "value": [1, 2]})

# CSV
df.to_csv("data.csv", index=False)
loaded = pd.read_csv("data.csv", usecols=["name"])

# Excel
df.to_excel("data.xlsx", sheet_name="Sheet1", index=False)
xls = pd.read_excel("data.xlsx", sheet_name="Sheet1")

# Parquet (columnar, compressed)
df.to_parquet("data.parquet")
pq = pd.read_parquet("data.parquet")

# SQL
engine = create_engine("sqlite:///app.db")
df.to_sql("items", engine, if_exists="replace", index=False)
from_db = pd.read_sql("SELECT * FROM items", engine)

Explanation

Pandas provides read_* and to_* helpers for many formats with sensible defaults. Parquet is columnar and compressed, ideal for large datasets, while CSV remains the lingua franca. to_sql and read_sql bridge DataFrames and databases through a SQLAlchemy engine.

More Pandas Snippets