Beyond the How: Why Table Partitioning Truly Matters
summary
Range based partitioning is presented as the only reliable approach: start partitioning around a single leading figure (transaction, policy, car) and align partition boundaries across all related tables, or joins will scan many partitions. Avoid hash, list and default partitions; prefer date or UUIDv7 keys so partitions prune predictably. Enable partitionwise join/aggregate after alignment to get real wins — the presenter reports up to 70x faster queries on systems tested at +100TB scale; begin thinking at ~100 GB table size.
AI-written summary · about
chapters (8)
- 0:00Introduction and passion
- 3:12Three reasons to partition
- 7:10Why range partitioning works
- 13:53Joining partitioned tables
- 21:50Leading figure concept
- 27:55Data lifecycle management
- 37:05Summary and key takeaways
- 42:49Q&A session
related talks
27mLeveraging table partitioning for query performance and data archiving
26mMy Postgres partitioning cookbook
31mPostgres Partitioning Best Practices
23mPostgres Partitioning Best Practices
30mPostgreSQL Partitioning: Slicing and Dicing for Performance and Easier Maintenance
26mPartitioning your Postgres tables for 20x better performance
45mIt's Not You, It's Me: Breaking Up with Massive Tables via Partitioning
30mOvercoming Performance Hurdles in Postgres Partitioned Tables