Constraints

This section explains how to manage constraints used for ensuring data integrity.

1. Types of constraint

The following constraint types are available:

Unique node property constraints

Unique property constraints ensure that property values are unique for all nodes with a specific label. Unique constraints do not mean that all nodes have to have a unique value for the properties — nodes without the property are not subject to this rule.

Node property existence constraints

Node property existence constraints ensure that a property exists for all nodes with a specific label. Queries that try to create new nodes of the specified label, but without this property, will fail. The same is true for queries that try to remove the mandatory property.

Relationship property existence constraints

Property existence constraints ensure that a property exists for all relationships with a specific type. All queries that try to create relationships of the specified type, but without this property, will fail. The same is true for queries that try to remove the mandatory property.

Node key constraints

Node key constraints ensure that, for a given label and set of properties:

  1. All the properties exist on all the nodes with that label.

  2. The combination of the property values is unique.

Queries attempting to do any of the following will fail:

  • Create new nodes without all the properties or where the combination of property values is not unique.

  • Remove one of the mandatory properties.

  • Update the properties so that the combination of property values is no longer unique.

Node key constraints, node property existence constraints and relationship property existence constraints are only available in Neo4j Enterprise Edition. Databases containing one of these constraint types cannot be opened using Neo4j Community Edition.

2. Implications on indexes

Creating a constraint has the following implications on indexes:

  • Adding a unique property constraint on a property will also add a single-property index on that property, so such an index cannot be added separately.

  • Adding a node key constraint for a set of properties will also add a composite index on those properties, so such an index cannot be added separately.

  • Cypher will use these indexes for lookups just like other indexes. Refer to Indexes for search performance for more details on indexes.

  • If a unique property constraint is dropped and the single-property index on the property is still required, the index will need to be created explicitly.

  • If a node key constraint is dropped and the composite-property index on the properties is still required, the index will need to be created explicitly.

Additionally, the following is true for constraints:

  • A given label can have multiple constraints, and unique and property existence constraints can be combined on the same property.

  • Adding constraints is an atomic operation that can take a while — all existing data has to be scanned before Neo4j can turn the constraint 'on'.

  • Best practice is to give the constraint a name when it is created. If the constraint is not explicitly named, it will get an auto-generated name.

  • The constraint name must be unique among both indexes and constraints.

  • Constraint creation is by default not idempotent, and an error will be thrown if you attempt to create the same constraint twice. Using the keyword IF NOT EXISTS makes the command idempotent, and no error will be thrown if you attempt to create the same constraint twice.