Compose with regular SQL

pgrdf.sparql(...) is a set-returning function. You can JOIN it to a Postgres table, use it as a CTE, funnel its result into INSERT INTO ... SELECT, or wrap it in a view.

What it does

Every UDF in the pgrdf.* schema is a regular Postgres function. That means every Postgres composition idiom works:

• pgrdf.sparql(...) in a FROM clause as a row source.
• WITH CTEs that name SPARQL results.
• JOIN SPARQL output against your application tables.
• INSERT INTO ... SELECT FROM pgrdf.sparql(...) to materialise.
• CREATE VIEW v AS SELECT ... FROM pgrdf.sparql(...).

The graph is just another table.

Why you'd use it

• Data scientists — keep your feature pipelines in SQL. Pull graph features in alongside relational ones in the same query.
• Application developers — no separate query language for graph vs. relational concerns. One ORM, one connection pool.
• Reporting / BI — wrap SPARQL queries as views, point your BI tool at them.

Example — JOIN graph data to your tables

WITH friends AS (
  SELECT * FROM pgrdf.sparql(
    'PREFIX foaf: <http://xmlns.com/foaf/0.1/>
     SELECT ?a ?b WHERE { ?a foaf:knows ?b }')
)
SELECT a.region, COUNT(*) AS knows_count
  FROM friends f
  JOIN app.users a ON a.iri = f."a"
 GROUP BY a.region
 ORDER BY knows_count DESC;

Example — materialise a SPARQL result as a table

CREATE TABLE app.person_emails AS
SELECT
    (s ->> 'p') AS person_iri,
    (s ->> 'm') AS mailbox
  FROM pgrdf.sparql(
    'PREFIX foaf: <http://xmlns.com/foaf/0.1/>
     SELECT ?p ?m WHERE { ?p foaf:mbox ?m }') AS s;

Example — wrap as a view

CREATE VIEW reports.engineers AS
SELECT
    (s ->> 'p') AS person,
    (s ->> 'n') AS name
  FROM pgrdf.sparql(
    'PREFIX ex: <http://example.com/>
     SELECT ?p ?n
       WHERE { ?p a ex:Engineer ; ex:name ?n }') AS s;

-- The view is now a first-class table-shaped object;
-- BI tools, ORMs, and grafana plug straight in.
SELECT * FROM reports.engineers ORDER BY name LIMIT 50;