Getting started

duckrouting adds graph routing to DuckDB. It runs the Boost Graph Library over any query that returns edges, and follows pgRouting's function semantics so existing routing SQL ports with little more than a rename.

Install

INSTALL duckrouting FROM community;
LOAD duckrouting;

To confirm the extension loaded, and see which Boost it was built against:

SELECT duckrouting_version('hello');
-- → Duckrouting hello, built against Boost 1_89

A first route

duckrouting does not need a topology table. Any relation with an id, two endpoints and a cost will do:

CREATE TABLE edges (
  id BIGINT, source BIGINT, target BIGINT, cost DOUBLE, reverse_cost DOUBLE
);
INSERT INTO edges VALUES
  (1, 5,  6, 1,  1), (2, 6, 10, -1, 1), (3, 10, 15, -1, 1),
  (4, 6,  7, 1,  1), (5, 10, 11, 1, -1), (6, 1,  3,  1, 1);

These edges are taken from pgRouting's sample network (CC BY-SA 3.0).

Then ask for a path:

SELECT node, edge, agg_cost
FROM duckrouting_dijkstra(
  'SELECT id, source, target, cost, reverse_cost FROM edges',
  5, 7);
-- → node  edge  agg_cost
-- →    5     1       0.0
-- →    6     4       1.0
-- →    7    -1       2.0

The first argument is a string containing SQL, not a table. That is how pgRouting works, and keeping it means the graph can be filtered, joined or computed on the fly without a separate materialised topology.

Where to go next

• The edges query -- the columns you must return, and how costs encode one-way streets.
• Dijkstra family -- the routing functions themselves.
• pgRouting function catalog -- what exists upstream, and which parts are Boost-backed.