
I built a Parquet toolkit that runs entirely in the browser — here's the architecture
The most common thing I do with a Parquet file is embarrassingly small: someone sends me...
Tag archive

The most common thing I do with a Parquet file is embarrassingly small: someone sends me...
How I run 90-day video analytics over Parquet exports with DuckDB — no warehouse, no server. Covers
How we moved cross-region video analytics off SQLite onto DuckDB querying partitioned Parquet export

A DuckDB community-extension PR sat red for eight weeks over a job named linux_arm64. The real bug was three words in the log— and a fleet of AI agent
How I run ad-hoc video analytics without a warehouse: export SQLite to partitioned Parquet, query it
The Data Engineering Holy Grail: Running intensive analytical queries without starving your...
How I run ad-hoc video analytics by exporting SQLite to Parquet and querying it with DuckDB — predic
Exporting trending-video metadata from production SQLite to hive-partitioned Parquet and running cro

A few things I noticed. None of them alarming on their own. Our database doubles in size every...
DuckDB is the closest thing the analytics world has to SQLite. It runs in-process, needs no server,...
TL;DR Traditional decoupled ETL pipelines (like the "Modern Data Stack") are too brittle...
How we replaced notebooks and a warehouse with DuckDB for ad-hoc video watch-time analytics on parti