pgGraph v0.1.0 — Graph queries for your existing Postgres data
Today we’re releasing the first public version of pgGraph, an open-source PostgreSQL extension that brings graph search, traversal, shortest path, and relationship queries directly to ordinary PostgreSQL tables.
No separate graph database.
No data pipeline to keep in sync.
No new source of truth.
Just Postgres.
Why pgGraph exists
Most application data already lives in PostgreSQL.
Users, accounts, companies, documents, payments, devices, organizations, permissions, events, they are all connected. But when you want to ask graph-shaped questions, you usually have to choose between:
- writing custom recursive SQL for every schema;
- copying data into a separate graph database;
- building a one-off indexing system yourself; or
- giving up and doing relationship exploration in application code.
pgGraph is an attempt to make those questions feel native to Postgres without huge performance penalty.
What you can do in v0.1.0
- Register existing PostgreSQL tables as graph nodes.
- Register relationships manually or auto-discover them from foreign keys.
- Build a derived graph index from relational data.
- Search graph nodes by property.
- Traverse relationships with bounded BFS and DFS.
- Find shortest paths between nodes.
- Run weighted shortest path queries (experimental).
- Compute connected components.
- Experiment locally with Docker, quickstart scripts, benchmarks, and a Streamlit playground.
Try it in 2 minutes
git clone https://github.com/evokoa/pggraph.git
cd pggraph
scripts/quickstart.sh
Find out more here:
https://docs.evokoa.com/pggraph