This directory contains examples demonstrating BenchBox's DataFrame benchmarking capabilities.

BenchBox supports native DataFrame benchmarking alongside traditional SQL database benchmarking. These examples show how to:

1. Run TPC-H/TPC-DS queries using Polars DataFrame API
2. Run queries using Pandas DataFrame API
3. Compare performance across platforms
4. Compare SQL vs DataFrame execution paradigms

Install the DataFrame platforms, then generate the benchmark data. The data is required before any example can execute.

pip install polars pandas

benchbox run --platform duckdb --benchmark tpch --scale 0.01 --phases load
benchbox run --platform duckdb --benchmark tpcds --scale 1 --phases load

Polars TPC-H (polars_tpch.py)

python examples/dataframe/polars_tpch.py

Demonstrates running TPC-H queries using Polars' expression API with lazy evaluation.

Pandas TPC-H (pandas_tpch.py)

python examples/dataframe/pandas_tpch.py

Demonstrates running TPC-H queries using Pandas' DataFrame API with eager evaluation.

Polars TPC-DS (polars_tpcds.py)

python examples/dataframe/polars_tpcds.py

Demonstrates running TPC-DS queries (more complex than TPC-H with 99 queries).

Cross-Platform Comparison (cross_platform_comparison.py)

python examples/dataframe/cross_platform_comparison.py

Compares query execution time across Polars and Pandas for the same queries.

SQL vs DataFrame (sql_vs_dataframe.py)

python examples/dataframe/sql_vs_dataframe.py

Compares SQL execution (DuckDB) vs DataFrame execution (Polars) for equivalent queries.

BenchBox uses a family-based architecture for DataFrame support:

result = (
    df.filter(col('shipdate') <= lit(cutoff))
    .group_by('status')
    .agg(col('quantity').sum().alias('total_qty'))
)

filtered = df[df['shipdate'] <= cutoff]
result = filtered.groupby('status').agg({'quantity': 'sum'})

For production benchmarking, use the BenchBox CLI:

benchbox run --platform polars-df --benchmark tpch --scale 0.1

benchbox run --platform pandas-df --benchmark tpch --scale 0.01

benchbox compare results_polars.json results_pandas.json

• DataFrame Platforms Overview
• Pandas DataFrame Platform
• TPC-H Guide
• TPC-DS Guide