From OLTP to OLAP: Is PostgreSQL Eating Analytics Too?
summary
Postgres with columnar extensions such as pg_duckdb, Citus and TigerData runs on EC2 and is compared to ClickHouse Cloud using ClickBench (104-column flat table, 43 single‑thread cold queries) and a 10 GB TPC-DS set (24 tables, 10 focused queries, 1,000 iterations, P95). Embedding DuckDB into Postgres yields concrete wins—short queries show improvements up to 7x faster, TPC‑DS gains near up to 3x, but some queries see regressions up to 2x and ClickHouse remains faster on many workloads. The extensions work but produce mixed, workload‑dependent results.
AI-written summary · about
chapters (10)
- 0:00Introduction to Postgres analytics
- 1:46Postgres in modern analytics stacks
- 3:29What defines analytical workloads
- 4:11Postgres extensions for analytics
- 8:23Performance benchmark objectives
- 9:26Benchmark setup details
- 11:31ClickBench performance results
- 13:58TPC-DS benchmark insights
- 15:45Summary and future research
- 17:29Q&A and closing remarks
related talks
49mCan PostgreSQL Compete with Analytics Specialists?
21mWhat can Postgres learn from DuckDB?
24mpg_duckdb: Ducking awesome analytics in Postgres
27mBuilding a PostgreSQL data warehouse
21mAnalytics in Postgres –A decade in the making
34mAnalytics and AI with PostgreSQL - Mission Possible!
46mPostgreSQL Analytics: It's a Trap!
34mAnalytics and AI with PostgreSQL - Mission Possible!