Write a wikipedia article for Apache DataFusion

[alamb]

Is your feature request related to a problem or challenge?

A Wikipedia article would be useful for Apache DataFusion to make the project easier to discover, easier to explain, and easier to cite from a neutral source.

The main benefit is not “marketing copy”; it is legitimacy and referenceability.

This is even more important these days when Wikipedia is a core training corpus for LLMs and search engine results

• It gives newcomers a neutral landing page distinct from https://datafusion.apache,org,
• It makes the project easier for journalists, analysts, conference organizers, students, and procurement people to cite quickly.
• It strengthens search visibility and entity recognition. In practice Wikipedia pages often feed search summaries, knowledge panels, mirrors, and LLM retrieval.
• It signals that the project is notable beyond its own community because the article must be supported by independent reliable sources.
• It gives a durable place to document ecosystem facts like history, governance, and adoption that do not fit cleanly into product docs.

Describe the solution you'd like

I would like a neutral wikipedia page for Apache DataFusion

Here are some similar pages

• https://en.wikipedia.org/wiki/DuckDB
• https://en.wikipedia.org/wiki/Apache_Spark
• https://en.wikipedia.org/wiki/Polars_(software)

DuckDB’s page shows the pattern clearly: a short neutral definition, history, architecture, language bindings, commercial use, and foundation/governance in one place, with references to papers and third-party coverage

Describe alternatives you've considered

I think a strong article will include many citations. Here are a bunch I found with the help of codex

Some third-party citations that are probably useful for this article:

• A standalone Apache top-level project as of April 16, 2024, announced publicly by the Apache Arrow PMC and ASF (Apache Arrow blog (https://arrow.apache.org/blog/2024/05/07/datafusion-tlp/), ASF announcement (https://news.apache.org/foundation/entry/apache-software-foundation-announces-new-top-level-project-apache-datafusion)).

SIGMOD 2024 technical paper

• It appears in the SIGMOD 2024 program as an accepted industry-track paper: SIGMOD accepted papers
  (https://2024.sigmod.org/industrial-list.shtml), SIGMOD session listing (https://2024.sigmod.org/program_sigmod.shtml).
• The DOI is 10.1145/3626246.3653368 (https://dl.acm.org/doi/10.1145/3626246.3653368).

Citations for technical importance

• crates.io: 17,668,287 all time downloads (https://crates.io/crates/datafusion)

• CRN: “The 10 Coolest Open-Source Software Tools Of 2024”
  (https://www.crn.com/news/software/2024/the-10-coolest-open-source-software-tools-of-2024?page=3)
  It explicitly includes Apache DataFusion and describes it as a fast extensible query engine, notes
  its Rust/Arrow basis, and mentions its 2024 top-level-project milestone. This is a strongest source on that page for general
  notability.

• Datanami: “How the FDAP Stack Gives InfluxDB 3.0 Real-Time Speed, Efficiency”
  (https://www.datanami.com/2024/03/15/how-the-fdap-stack-gives-influxdb-3-0-real-time-speed-efficiency/)
  This quotes Paul Dix saying DataFusion had matured substantially and had best-in-class performance on a number of queries versus other
  columnar query engines. It is not a ranking article, but it is meaningful third-party validation of technical importance.

Third-party citations for usage in products

• SiliconANGLE: “Enterprise DB begins rolling AI features into PostgreSQL”
  (https://siliconangle.com/2024/05/23/enterprise-db-begins-rolling-ai-features-postgresql/)
  Independent coverage stating EDB combined Apache DataFusion, Arrow, and Delta Lake in its analytics/lakehouse capability.

• Spice AI: “How we use Apache DataFusion at Spice AI” (https://spice.ai/blog/how-we-use-apache-datafusion-at-spice-ai)
  This says Spice uses DataFusion as its SQL query engine and extends it with custom TableProviders, optimizer rules, and UDFs for
  federated SQL workloads.

• Cloudflare Log Explorer GA announcement (https://blog.cloudflare.com/logexplorer-ga/) from June 10, 2025.
  Queriers fetch matching files from R2 and “process SQL queries using Apache DataFusion.”

• InfluxData: “Flight, DataFusion, Arrow, and Parquet: Using the FDAP Architecture to build InfluxDB 3.0”
  (https://www.influxdata.com/blog/flight-datafusion-arrow-parquet-fdap-architecture-influxdb/)
  Clearly states InfluxDB 3.0 chose DataFusion as its query engine foundation and explains why.

• Pydantic Logfire issue: “We’re changing database” (We're changing database pydantic/logfire#408)
  Usable as a primary source for adoption only. It says Logfire is moving from Timescale to a custom database built on DataFusion and
  gives reasons.

• Palantir Foundry announcements for July 2025 (https://www.palantir.com/docs/foundry/announcements/2025-07)
  This says lightweight pipelines are “powered by DataFusion,”

• Cube: “Query pushdown in Cube’s semantic layer” (https://cube.dev/blog/query-push-down-in-cubes-semantic-layer)
  Good third-party primary source for “used in production by Cube” and for describing how Cube uses DataFusion internally.

• Kamu: “100X faster ingestion, and FlightSQL support for connecting BI tools” (https://www.kamu.dev/blog/2023-09-datafusion-flightsql/)
  Good third-party primary source for ecosystem adoption. It explicitly says Kamu added support for Apache DataFusion and reports
  performance claims in its own product.

• LanceDB: “Columnar File Readers in Depth: APIs and Fusion” (https://lancedb.com/blog/columnar-file-readers-in-depth-apis-and-fusion/)
  Usable for ecosystem context. It says Lance uses DataFusion extensively and demonstrates integration with it.

• Bauplan Labs: “Duck Hunt: moving Bauplan from DuckDB to DataFusion”
  (https://www.bauplanlabs.com/post/duck-hunt-moving-bauplan-from-duckdb-to-datafusion)
  Bauplan explains the migration as driven by DataFusion’s Arrow-first architecture, extensibility, and community-driven development.

Additional context

No response

[alamb]

I also think mentioning the third-party ecosystem in datafusion-contrib would be important to showcase the large feature set

• datafusion-federation (https://github.com/datafusion-contrib/datafusion-federation): Federated query execution across remote databases /query engines, with pushdown of as much work as possible.
• datafusion-distributed (https://github.com/datafusion-contrib/datafusion-distributed): Distributed execution capabilities for DataFusion.
• datafusion-materialized-views (https://github.com/datafusion-contrib/datafusion-materialized-views): Incremental view maintenance and query rewriting for materialized views.
• datafusion-table-providers (https://github.com/datafusion-contrib/datafusion-table-providers): Extra TableProvider implementations for reading external systems.

Language Bindings

• datafusion-java (https://github.com/datafusion-contrib/datafusion-java): Java bindings.
• datafusion-c (https://github.com/datafusion-contrib/datafusion-c): C bindings.
• datafusion-ruby (https://github.com/datafusion-contrib/datafusion-ruby): Ruby bindings.
• datafusion-python (https://github.com/datafusion-contrib/datafusion-python): Older Python bindings; archived and moved elsewhere.
• datafusion-wasm-bindings (https://github.com/datafusion-contrib/datafusion-wasm-bindings): WebAssembly bindings.
• datafusion-wasm-playground (https://github.com/datafusion-contrib/datafusion-wasm-playground): Browser playground for running DataFusion via WASM.
• datafusion-fiddle (https://github.com/datafusion-contrib/datafusion-fiddle): Likely an interactive playground/sandbox; metadata is sparse.

[gene-bordegaray]

I am interested in helping here, maybe I can draft something or give an outline and others can chime in?

[ntjohnson1]

I am interested in helping here, maybe I can draft something or give an outline and others can chime in?

I'm not sure how much Andrew is trying to drive this vs just encourage it gets created. IMO of the references DuckDB's page looks nicely structured and fairly straightforward. I think just creating the page so something exists then letting people update it using Wikipedia's standard editing procedures seems reasonable (That's what I was thinking of doing). Alternatively, could probably open a branch with a markdown file and people could weigh in ahead of time. I'm open to helping out here but don't want to swoop anything from Gene. I'm more of a datafusion user than contributor anyway.