We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
Hacker News·Jun 24·5 min read
I rewrote PostHog's SQL parser, 70x faster, while barely looking at the code
TL;DR
PostHog needed a more efficient SQL parser to improve query performance. A new parser was built using ANTLR and Claude Code, resulting in a 70x speed increase.
✦ Why It Matters
Engineers can leverage AI tools like Claude Code to significantly enhance software performance and efficiency.
Key Takeaways
- Explore using ANTLR for building custom parsers.
- Consider AI tools for optimizing code performance.
- Test edge cases thoroughly to ensure compatibility.
- Measure performance improvements quantitatively.
Full Summary
PostHog allows users to query their data directly using SQL, which is then converted to raw ClickHouse SQL for processing. The original SQL parser was inefficient, prompting a rewrite using ANTLR, a powerful parser generator, and Claude Code, an AI coding assistant.
The new implementation consists of 16,000 lines of custom parser code, 5,000 lines of supporting tools, and additional test cases. The result was a remarkable 70x speed improvement while ensuring that the parser remains functionally equivalent for all realistic SQL queries.
Only a small subset of edge cases, such as repeated keywords, were not supported. This project demonstrates the potential of leveraging AI tools for significant performance enhancements in software development.