```{toctree}
---
maxdepth: 2
hidden: true
---



docs/tutorial.md

docs/how-to-integrate.md

docs/how-to-chain-filesystems.md

docs/explanation.md

docs/api.md


```

# fsspec-data

Arrow interchange planning for the fsspec data ecosystem.

[![Build Status](https://github.com/1kbgz/fsspec-data/actions/workflows/build.yaml/badge.svg?branch=main&event=push)](https://github.com/1kbgz/fsspec-data/actions/workflows/build.yaml)
[![codecov](https://codecov.io/gh/1kbgz/fsspec-data/branch/main/graph/badge.svg)](https://codecov.io/gh/1kbgz/fsspec-data)
[![License](https://img.shields.io/github/license/1kbgz/fsspec-data)](https://github.com/1kbgz/fsspec-data)
[![PyPI](https://img.shields.io/pypi/v/fsspec-data.svg)](https://pypi.python.org/pypi/fsspec-data)

`fsspec-data` reconciles provided and requested tabular schemas and formats without taking
ownership of database discovery or dataframe query planning. A pure-Rust core and Python
bindings apply the same decisions to Arrow record batches and PyArrow objects.

The first release supports format-aware conversion plans, exact/projection/compatible/
coerce schema policies, registered Arrow IPC/Parquet/CSV/JSONL codecs, bounded lazy batch
iteration with cancellation, and read-only conversion through chained fsspec URLs.

## Documentation

- [Convert tabular data](docs/tutorial.md) in a
  guided first example.
- [Connect a batch producer or consumer](docs/how-to-integrate.md)
  to the interchange layer.
- [Convert a file through an fsspec chain](docs/how-to-chain-filesystems.md)
  while retaining its source filesystem.
- [Understand the interchange design](docs/explanation.md)
  and package boundaries.
- [Look up the interchange API](docs/api.md),
  including policies, codecs, and errors.