Teide

Fast columnar dataframe library for Python, powered by a zero-dependency C17 engine with lazy fusion execution.

Install

pip install teide

Note: v0.1.0 supports Linux and macOS only. Windows support is planned.

Quick Start

from teide import Context, col

with Context() as ctx:
    df = ctx.read_csv("data.csv")

    result = (
        df.filter(col("price") > 0)
          .group_by("category")
          .agg(col("price").sum(), col("price").mean())
          .sort("price_sum", descending=True)
          .collect()
    )

    print(result)

Features

• Lazy evaluation with automatic query optimization (predicate pushdown, CSE, operator fusion)
• Morsel-driven parallel execution across all cores
• Zero-copy NumPy interop for numeric columns
• Fast CSV reader — parallel parsing, mmap I/O
• Columnar storage — splayed tables, date-partitioned datasets
• Pure C17 engine — no external dependencies, minimal memory overhead

License

MIT