Apache Doris — Hackathon at Community over Code Glasgow 2026

Apache Community over Code 2026 — Glasgow, UK, October 11–14

Register now | Event website | Hackathon overview

Apache Doris — Hackathon

Apache Doris is an open-source, real-time database for modern analytics and AI. It answers sub-second queries under high concurrency at petabyte scale; it treats structured, text and vector data as first-class citizens in the same engine, so a single SQL statement can combine a filter, a BM25 text match and a nearest-neighbour search over embeddings; it queries Iceberg, Delta Lake and Hudi tables in place, with no copy, which makes it an acceleration layer over an open lakehouse rather than yet another silo; and increasingly it serves as the store agents run on — columnar JSON for tool-call traces, BM25 retrieval for RAG and agent memory, MCP for access.

That range is the reason Doris is an interesting codebase to spend an afternoon in. Query execution, index structures, storage formats, connectors and SQL semantics all live in one repository, so whichever of those you find interesting, there is a task here that touches it.

Session: Tuesday 13 October, Early slot (11:20–15:00), Wee Dram Room

Coordinator

What We’re Working On

We’ll bring tasks in five areas. A curated task list with specific issues will be published on this page closer to the event; in the meantime, the directions below should tell you whether there’s something here for you.

Resources

Getting Started Before the Event

What to prepare depends on which direction you pick:

  1. If you plan to change Doris — get a build working ahead of time using the official development Docker image. Doris is a large C++ and Java codebase, a first full build takes a while, and building inside the image saves you from assembling the third-party dependencies by hand. Start this the day before at the latest, not on the morning of the hackathon.
  2. If you plan to build something on Doris — you don’t need to compile anything. Follow the quick start to get a cluster running locally, and bring whatever language and client you’re comfortable with.
  3. Decide roughly which direction interests you from the five above, so we can point you at a matching task on the day.
  4. Say hello on Slack or the dev@ list so we know to look out for you.

If none of that works out, come anyway — documentation needs nothing but a text editor, and we can sort out builds and clusters in the room.