The video for my "What Goes Around Comes Around... And Around" talk at CWI DA is now available: youtube.com/watch?v=8Woy5I… 📊Slides: cs.cmu.edu/~pavlo/slides/… 📄Paper: db.cs.cmu.edu/papers/2024/wh…

Some observations, open source projects that backed by company (that their product is the open source project) are much less responsive to outside pull requests

The latest paper from the #1 CMU-DB PhD student Sam Arch 🇦🇺's is wild compilation DB magic! He automatically makes UDFs run 300x faster on Microsoft SQL Server and 1.3x faster on DuckDB. Code: github.com/SamArch27/PRISM Paper: vldb.org/pvldb/vol18/p1…

Buckle up because we're crashing into the new year with my annual database retrospective: License change blowbacks! Databricks vs. SnowflakeDB gangwar! @DuckDB shotgun weddings! Buying a college quarterback with database money for your new lover! cs.cmu.edu/~pavlo/blog/20…

I really appreciate Andrew Lamb, and how he interacts with other devs in the ApacheArrow and ApacheDataFusion repo

I wish I could work on open source full time

New CMU Database Group Course: Database Query Optimization! A survey of how to build a SQL optimizer from scratch covering foundational and state-of-the-art implementations. All lectures available on Youtube: 15799.courses.cs.cmu.edu/spring2025/

I just LOVE contributing to open source If you use some tool and see a way to improve it, just open a PR. Maintaining one is hard

Best line ever

DataFusion will finally have tree style explain plans: github.com/apache/datafus… (likely in 47.0.0)

As soon as I see this I immediately don't trust the channel. If they don't do the work and just update the title and thumbnail, they probably did not do the work to research properly

I'm impressed by Cognition DeepWiki. It has ingested Fastify and pino code and created some explanations that are *better* than ours. The key feature for me is sequence diagrams.

I remember when Liran used to post things about security

Mutation testing is the best testing strategy and more should use it. I have seen bad code with bugs and even more bad tests, this will help those

No. Next question.

Mutli-level merge sort queued up for DataFusion 50.0.0 next month: github.com/apache/datafus… Thanks to Raz Luvaton ⬢ and Yongting You