BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//sf26-agenda//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
UID:SESS-124@sf26-agenda
DTSTAMP:20261020T151709Z
DTSTART:20261104T220000Z
DTEND:20261104T224500Z
SUMMARY:Hardwood: Building a Parquet Parser From Scratch (With a Little Hel
 p From AI)
LOCATION:Room 2006
DESCRIPTION:Gunnar Morling\n\nApache Parquet has become the lingua franca o
 f the modern data ecosystem\, powering data lakes and table formats like A
 pache Iceberg—but for Java\, the go-to library parquet-java pulls in Had
 oop and a truckload of other dependencies\, and its reader is single-threa
 ded. This was bugging me enough to start Hardwood\, a brand-new Parquet pa
 rser written from scratch in modern Java\, applying some of the performanc
 e lessons learned from the One Billion Row Challenge. Come and join me for
  this session\, where we'll look at: The internals of the Parquet format a
 nd what makes parallelizing its decoding surprisingly tricky Techniques fo
 r achieving high throughput\, such as page-level parallelism\, adaptive pr
 efetching\, and avoiding auto-boxing in hot loops How to use JDK Flight Re
 corder for identifying performance bottlenecks Practical learnings from us
 ing AI (specifically\, Claude Code) as a coding companion—what works wel
 l\, where you need to stay sharp\, and why "built with AI" doesn't mean "v
 ibe-coded" Whether you're interested in file formats\, Java performance\, 
 or getting a realistic take on AI-assisted development\, there should be s
 omething in here for you.
URL:https://sf26.rmoff.net/s/hardwood-building-a-parquet-parser-from-scratc
 h-with-a-124
END:VEVENT
END:VCALENDAR