Skip to main content

Indexes & Constraints

Property Indexes​

Indexes speed up MATCH queries that filter on specific properties.

Create Index​

Both modern (Neo4j 4.x/5.x) and classic (openCypher 9) syntax are supported:

-- Modern syntax (recommended)
CREATE INDEX person_name IF NOT EXISTS FOR (p:Person) ON (p.name)

-- Classic syntax
CREATE INDEX ON :Person(name)

After creating this index, queries like MATCH (p:Person {name: 'Alice'}) will use the index instead of scanning all Person nodes.

Composite Index​

Index multiple properties together:

-- Modern syntax
CREATE INDEX person_name_age IF NOT EXISTS FOR (p:Person) ON (p.name, p.age)

-- Classic syntax
CREATE INDEX ON :Person(name, age)

Drop Index​

-- Modern syntax (by name)
DROP INDEX person_name IF EXISTS

-- Classic syntax
DROP INDEX ON :Person(name)

Show Indexes​

SHOW INDEXES

Unique Constraints​

Constraints enforce data rules at the database level.

Create Unique Constraint​

-- Modern syntax (recommended)
CREATE CONSTRAINT person_email IF NOT EXISTS FOR (p:Person) REQUIRE p.email IS UNIQUE

-- Classic syntax
CREATE CONSTRAINT ON (p:Person) ASSERT p.email IS UNIQUE

After this, attempting to create two Person nodes with the same email will fail.

Create NOT NULL Constraint​

CREATE CONSTRAINT person_name_required IF NOT EXISTS FOR (p:Person) REQUIRE p.name IS NOT NULL

Drop Constraint​

DROP CONSTRAINT person_email IF EXISTS

Show Constraints​

SHOW CONSTRAINTS

Vector Indexes​

For similarity search on embedding vectors.

Create Vector Index​

CREATE VECTOR INDEX product_embed
  FOR (p:Product) ON (p.embedding)
  OPTIONS {dimensions: 384, similarity: 'cosine'}

Options:

  • dimensions (required): Vector dimensionality
  • similarity: 'cosine' (default), 'euclidean', or 'dot_product'

Query Vector Index​

CALL db.index.vector.queryNodes('Product', 'embedding', [0.1, 0.2, ...], 10)
YIELD node, score
RETURN node.name, score

Schema Introspection​

-- All labels in use
SHOW LABELS

-- All relationship types
SHOW RELATIONSHIP TYPES

-- All property keys
SHOW PROPERTY KEYS

-- Visual schema (returns nodes/edges representing the schema)
CALL db.schema.visualization()

Index Usage in Queries​

The query planner automatically uses indexes when available:

-- With index on :Person(name), this uses IndexScan instead of NodeScan
MATCH (p:Person {name: 'Alice'}) RETURN p

-- Check with EXPLAIN
EXPLAIN MATCH (p:Person {name: 'Alice'}) RETURN p
-- Output shows: IndexScan instead of NodeScan+Filter