pg_duckdb: Postgres analytics just got faster with DuckDB

Postgres analytics 10x faster with just an extension?! 🤯

In August, we announced pg_duckdb, a PostgreSQL extension that integrates DuckDB's analytics engine directly into Postgres. It's open-source and represents a joint partnership between Hydra and MotherDuck.

Two months later, we are happy to announce its first release and highlight many features, including the ability to read and write over object storage with Parquet and CSV, as well as Apache Iceberg (currently read-only), and the capability to query from MotherDuck without leaving Postgres.

Note :
We ingested the TPC-DS datasets into PostgreSQL without indexes for two main reasons:
1. Currently, pg_duckdb does not support indexes, which makes a direct comparison impossible. Addressing this limitation is a high priority for us (see GH issue : https://github.com/duckdb/pg_duckdb/i...)
2. While indexes are common in real-world PostgreSQL scenarios, optimizing them for specific analytic queries can be complicated and bring extra overhead. Considering this, we believe there is value in looking at the performance of queries without any indexes.

☁️🦆 Start using DuckDB in the Cloud for FREE with MotherDuck : https://hubs.la/Q02QnFR40

📓 Resources
Github Repository of pg_duckdb: https://github.com/duckdb/pg_duckdb
Blog announcement: https://motherduck.com/blog/pgduckdb-...

➡️ Follow Us
LinkedIn:   / motherduck  
X/Twitter :   / motherduck  
Blog: https://motherduck.com/blog/

0:00 Intro
1:33 Postgres extension ecosystem
2:35 Getting started with pg_duckdb
6:20 Query data lake / lakehouse
8:54 Scaling to the cloud with MotherDuck
13:37 Moving forward