akimbo

For when your data won't fit in your dataframe

Akimbo provides fast vectorized processing of nested, ragged data in dataframes, using the .ak accessor.

Features

• numpy-like API for slicing and accessing data deep in nested structures
• apply any function, ufunc or aggregation at any level
• use with different backends: pandas, polars, dask-dataframe and cuDF exactly the same way (more backends may come in the future)
• CPU and GPU processing and support for numba-jit
• attach object-like behaviours to your record (sub)structures

See the quick start in the documentation for an introduction to akimbo.

Acknowledgements

Support for this work was provided by NSF grant OAC-2103945.