DataFusion Community Showcase Vol.1

​We are excited to present the inaugural DataFusion Community Showcase!

​This first virtual session will highlight two projects building on DataFusion:

​Dewey Dunnington (Wherobots / Apache Sedona):

​In September, 2025, the Apache Sedona community released SedonaDB, a purpose-built engine based on DataFusion. The project consists of (1) purpose-built extensions that map spatial concepts to relational database concepts that can be optimized and executed in parallel by the underlying DataFusion engine and (2) a purpose-built Python interface that handles spatial Python objects and is specifically designed to interoperate with existing Python-based spatial workflows. Since its initial release we have built SedonaDB into a fully-featured spatial engine including support for GDAL read/write, larger-than memory spatial joins, cloud-native format support, and over 200 spatial functions with high PostGIS compatibility.

​Alex Merose (Xarray-SQL):

​Did you know that any Nd array can be represented as a 2d table? We exploit this fact to transform Xarray datasets into a tabular representation that can be queries with SQL: This is the Python package Xarray-SQL. We find resonate affordances between Apache DataFusion, Apache Arrow, and Xarray that maintain lazy evaluation and efficient parallelism. Thus, this humble library has the potential to make sense of planetary scale data (TiB - PiB range queries) in a way that's simple, declarative, and fault-tolerant. In this talk, I'll use Xarray-SQL as the basis to argue that all common geospatial and climate operations are actually relational operations in disguise.

​If you are interested in presenting at a future Community Showcase, please sign up here. And for more info, check out the DataFusion repo.