SQL Overview

NodeDB uses SQL as its primary query language. Whether you connect via ndb, psql, or HTTP — the same SQL works everywhere.

Query Execution

ndb CLI (NDB protocol)   ──┐
psql    (pgwire)         ──┼──► SQL Parser ──► EngineRules ──► PhysicalPlan ──► Data Plane
HTTP    (REST/JSON)      ──┘

Three doors, one room. Same parser, same optimizer, same execution engine.

Standard SQL Support

NodeDB supports standard SQL with high PostgreSQL compatibility:

• SELECT with WHERE, GROUP BY, HAVING, ORDER BY, LIMIT, OFFSET
• JOINs (INNER, LEFT, RIGHT, CROSS, SEMI, ANTI)
• Window functions (ROW_NUMBER, RANK, LAG, LEAD, SUM OVER)
• CTEs (WITH, WITH RECURSIVE)
• Subqueries and derived tables
• Set operations (UNION, INTERSECT, EXCEPT)
• Transactions with savepoints (BEGIN, COMMIT, ROLLBACK, SAVEPOINT)
• Prepared statements

Engine-Specific Extensions

Beyond standard SQL, NodeDB extends the language for each engine:

• Vector — SEARCH ... USING VECTOR(), vector_distance()
• Graph — GRAPH TRAVERSE, GRAPH ALGO, MATCH patterns
• Full-Text — text_match(), bm25_score(), search_score()
• Spatial — ST_DWithin(), ST_Contains(), ST_Distance(), etc.
• Timeseries — time_bucket(), ts_rate(), continuous aggregates
• KV — KV_INCR(), KV_CAS(), sorted indexes, rate gates
• CRDT — crdt_state(), crdt_apply()

Isolation Level

Snapshot Isolation (SI). Reads see a consistent snapshot from BEGIN time. Write conflicts detected at COMMIT.