Better sql for "Neo4j v.s. MySQL" in http://www.slideshare.net/slidarko/graph-windycitydb2010

graph_traversal.sql

DROP TABLE IF EXISTS t;
CREATE TABLE t (
	id	int PRIMARY KEY
	,adj	int[]
);

CREATE INDEX _idx_adj on t USING gin (adj);

INSERT INTO t VALUES
	(1, ARRAY[2,3,4,5])
	,(2, ARRAY[1])
	,(3, ARRAY[1,6])
	,(4, ARRAY[1])
	,(5, ARRAY[1,7])
	,(6, ARRAY[3])
	,(7, ARRAY[5]);

WITH RECURSIVE r (id, adj, step) AS (
	SELECT
		*
		,1
	FROM
		t
	WHERE
		id = 1
	UNION ALL
	SELECT
		t.*
		,r.step + 1
	FROM
		t,r
	WHERE
		t.id = ANY(r.adj) AND step < 5
)
SELECT
	*
FROM
	r
WHERE
	step = 5
ORDER BY
	id
;

Hey,

Have you tried the tests with this improved schema ? is there any notable difference in the benchmarks ?