Convert tabular formats and reconcile Arrow schemas over fsspec.
fsspec-data plans schema mappings and executes them over Arrow record batches.
Codecs handle Arrow IPC, Parquet, CSV, and JSONL. A read-only chained filesystem
exposes converted files; a separate converter registry handles resources such as
Xarray datasets and Zarr stores.
Convert CSV identifiers to Parquet walks through planning, batch iteration, and encoded output.
For application work, connect a batch producer or consumer, convert a file through an fsspec chain, or resolve schemas from files and registries. Convert Xarray and Zarr resources covers native resource routes and converter registration.
The interchange API reference covers codecs, policies, limits, and planning. Schema and converter reference describes resolution and resource routes. Read Interchange design for the boundaries between schemas, data, and consumers.