README.md

datafusion-extra-functions

Extra aggregate functions for Apache DataFusion. This is not an official Apache Software Foundation release.

Version 0.5.5 supports DataFusion 55.0 and Rust edition 2024.

Installation

cargo add datafusion-extra-functions

From Python (datafusion-python 55), install the FFI wheel and register aggregates via the PyCapsule protocol. See python/README.md.

pip install datafusion-extra-functions

Register all functions with your session context:

datafusion_extra_functions::register_all_extra_functions(&mut ctx)?;

Available functions

Function	Description
mode(expression)	Returns most frequent value.
max_by(value, key)	Returns value at maximum key.
min_by(value, key)	Returns value at minimum key.
skewness(expression)	Computes skewness.
kurtosis_pop(expression)	Computes population excess kurtosis.
kurtosis(expression)	Computes sample excess kurtosis.

SELECT mode(city) FROM visits;
SELECT max_by(user_id, score) FROM results;
SELECT min_by(user_id, score) FROM results;
SELECT skewness(measurement) FROM readings;
SELECT kurtosis_pop(measurement) FROM readings;
SELECT kurtosis(measurement) FROM readings;

Testing

cargo test

Python bindings (from python/):

pip install maturin pytest
maturin develop --release
pytest