Walking trees with SQL

Adjacency list

CREATE TABLE taxon (
  id INT PRIMARY KEY,
  parent_id INT NULL,
  name VARCHAR(100),
  KEY (parent_id)
);

Cheap writes. Deep “all descendants” queries hurt without recursion helpers.

Nested sets / path enumeration

Fast subtree reads. Expensive writes. Choose when reads dominate.

OQGRAPH

Keep edges in a backing table; query algorithms in SQL (docs):

SELECT GROUP_CONCAT(linkid ORDER BY seq) AS lineage
FROM oq_graph
WHERE latch='dijkstras' AND origid=:root AND destid=:leaf;

Choose

Pattern	Reads	Writes
Adjacency	slower deep walks	cheap
Nested sets / paths	fast subtrees	expensive
OQGRAPH	path/reachability	normal edge DML

Match the traversal you run — not a universal tree religion.