Overview#

FeatherStore is high performance datastore for storing Pandas DataFrames, Polars DataFrames, and PyArrow Tables. By saving data in the form of partitioned Feather Files, FeatherStore enables several operations on the stored tables, optimizing performance by selectively loading only the necessary segments of data:

  • Partial reading of data

  • Append data

  • Insert rows and columns (Pandas, Polars DataFrame, or PyArrow Table)

  • Update data (Pandas, Polars DataFrame, or PyArrow Table)

  • Drop data

  • Read metadata (including column names, indices, table dimensions, etc.)

  • Changing column types

For more information on using FeatherStore, please refer to the user guide.

Installation#

The project is hosted on PyPI at:
https://pypi.org/project/FeatherStore/
To install FeatherStore, simply use pip
pip install featherstore
or
pip install git+https://github.com/hakonmh/featherstore.git
to install the latest version from GitHub.

Python version support#

Officially Python 3.11 and up is supported.

Dependency requirements#

FeatherStore 0.3.0 requires:

  • pandas >= 2.2.0

  • polars[timezone] >= 1.21.0

  • pyarrow >= 14.0.0

These are installed automatically with pip install featherstore.

Source Code#

The source code is currently hosted on GitHub at:
Hakonmh/featherstore

LICENSE#

MIT

Contributions#

All contributions, bug reports, bug fixes, documentation improvements, enhancements and ideas are welcome.

Issues are posted on:
hakonmh/featherstore#issues