FANTASTIC DATABASES AND WHERE TO FIND THEM
summary
Compares concrete database choices and tricks: SQLite powers sites doing 20 million queries a day on shared VMs, gives easy file backups and fast reads but has single-writer locks; Postgres uses unlogged tables to cut CI test time (examples go from ~3 minutes to ~20 seconds), window functions and cumulative sums express analytics in one query via ActiveRecord, and explain buffers reveal planning vs execution and cache hits; index includes and partial indexes avoid table hits; columnar and distributed extensions and ClickHouse (Cloudflare 6M req/sec) show when to shard or compress.
AI-written summary · about
chapters (12)
- 0:00Mythical database intro
- 3:58Why databases feel mythical
- 5:45SQLite hidden superpower
- 6:50Postgres Swiss Army knife
- 8:12Postgres for Ruby devs
- 10:38Postgres query execution
- 15:09Optimizing Postgres indexes
- 17:35Postgres extensions storage
- 19:22Distributed databases sharding
- 20:24Exotic databases overview
- 21:46When to use NoSQL
- 22:26Closing thoughts
related talks
41mUsing Databases to Pull Your Application Weight
45mDatabases for Data Science
41mHow to choose the right database for your application
43mJust Use Postgres.
34mSimple and Awesome Database Tricks
30mUse SQLite in production
31mUsing Databases to pull your applications weight
35mAn Ode to 17 Databases in 39 minutes