Columnar Databases Are Incredibly Fast. Here's Why.

🔗 Relevant Links Postgres is not the best at everything. The same SQL on the same 100 million rows runs over 40 times faster in ClickHouse and DuckDB, then falls apart on a single row update. Columnar databases store columns together instead of rows, and that one choice explains both results. If you run analytics on Postgres, watch this before you add another index. ❤️ More about us Radically better observability stack: https://betterstack.com/ Written tutorials: https://betterstack.com/community/ Example projects: https://github.com/BetterStackHQ 📱 Socials Twitter: https://twitter.com/betterstackhq Instagram: https://www.instagram.com/betterstackhq/ TikTok: https://www.tiktok.com/@betterstack LinkedIn: https://www.linkedin.com/company/betterstack 📌 Chapters: 00:00 Postgres is not the best at everything 00:43 Group by on 100 million rows: 40x faster 01:45 Count events in a date range 02:14 How ClickHouse and DuckDB skip most of the disk 03:14 Select one row by ID: Postgres wins 04:20 Update one row: ClickHouse rewrites the column 05:09 Count distinct users 05:39 ClickHouse vs DuckDB: server or library 06:09 Which one should you pick