# parquet-proxy-odata

A proxy that allows you to query OData API's as Parquet with accurate type conversion and basic caching.

## Getting started
```bash
# Install dependencies
uv sync

# Start the proxy
uv run fastapi dev parquet-proxy-odata/app.py

# Use the proxy
duckdb <<EOF
set force_download=True;
select * from read_parquet("http://127.0.0.1:8000/datasets?url=https://opendata.cbs.nl/ODataApi/odata/85828NED/TypedDataSet");
EOF
```