GraphAcademy

Cypher Indexes and Constraints

Indexes and Constraints in Neo4j

Using Constraints in Neo4j

Using Indexes in Neo4j

Using Full-text Indexes in Neo4j

Index Usage in Neo4j

Indexes in Neo4j

Indexes in Neo4j

An index in Neo4j is a data structure that allows the graph engine to retrieve data quickly. All indexes in Neo4j require more storage in the graph, so you must ensure that you do not index everything!

You learned in the previous module that a best practice is to create uniqueness constraints before you load your data. After the data is loaded, you create additional indexes to make your queries perform faster. Using indexes makes writing data slower, but retrieving it faster.

Uniqueness constraints are implemented as indexes, but there are more types of indexes that you can create and use:

  • RANGE

  • Composite

  • TEXT

  • POINT

  • Full-text

Full-text indexes are used differently from other indexes and will be covered in the next module.

This course does not currently cover POINT indexes.

Using PROFILE to analyze queries that use indexes

Previously in this course, you learned about the importance of creating the required indexes for the use cases of your application. At runtime, only one index is used by default. This means that not only must you plan what indexes are appropriate for your use cases, but also that the amount of data in the graph will impact what indexes are used at runtime.

In this module you will PROFILE queries to:

  • Identify which index is used for a query.

  • Determine if adding an index improves query performance.

RANGE indexes

A b-tree is a common implementation of an index that enables you to sort values. A RANGE index in Neo4j is a proprietary implementation of a b-tree. You can define a RANGE index on a property of a node label or relationship type. The data stored in the index can be any type.

A RANGE index can speed up the following in your Cypher code:

  • Equality checks =.

    • Note: For string properties, a TEXT index may perform better.

  • Range comparisons >,>=,<, <=

  • STARTS WITH string comparisons

    • Note: TEXT indexes can be used for STARTS WITH comparisons but may not perform as well as RANGE indexes.

  • Existence checks IS NOT NULL

ENDS WITH ands CONTAINS comparisons may benefit slightly from a RANGE index, but we recommend you use a TEXT index for these types of comparisons.

This query could benefit from a RANGE index in the graph for the born property associated with the Person label:

cypher
MATCH (p:Person)
WHERE p.born IS NOT NULL
RETURN p

Composite indexes

A composite index combines values from multiple properties for a node label or for relationship type. You create composite indexes when multiple properties are always tested together in a query. The types of the properties need not be the same.

During query planning, at most one index is used, so it is beneficial in some cases to create a composite index when multiple properties need to be retrieved quickly.

For example, you might want to index on the Movie.year and the Movie.title properties. If you created a composite index on these two properties, then this query would perform better:

cypher
MATCH (m:Movie) WHERE m.year > 1999
AND m.title CONTAINS "Toy"
RETURN m.title, m.year, m.imdbRating

TEXT indexes

A TEXT index supports node or relationship property types that must be strings.

A TEXT index can speed up the following in your Cypher code:

  • Equality checks =

  • String comparisons ENDS WITH, CONTAINS

  • List membership x.prop in ["a","b","c"]

About LOOKUP indexes

There are two types of LOOKUP indexes in a Neo4j database. These indexes are created automatically for you and are used to look up a node label or relationship type value.

Without these indexes queries would need to scan the entire set of nodes or relationships for some queries which is very inefficient.

You should not create or remove these types of indexes in the database as they are automatically created for you.

Index not used

A predicate such as the following does use a TEXT or RANGE index:

cypher
MATCH (p:Person)-[:ACTED_IN]->(m:Movie)<-[:DIRECTED]-(p2:Person)
WHERE p.name = p2.name
RETURN p

POINT indexes

This course does not currently cover POINT indexes. You can read about them in the Neo4j Cypher Manual.

Check your understanding

1. Indexes for numeric values

What types of indexes support numeric values? (Select all that apply.)

  • ✓ RANGE

  • ❏ TEXT

  • ❏ full-text

  • ✓ Composite

Hint

There are two types of indexes that can be used to test numeric property values.

Solution

RANGE and Composite indexes can support properties with numeric values.

2. What type of index possible?

Suppose you need to create an index on the Movie.title property to speed up this query:

cypher
MATCH (m:Movie)
WHERE m.title CONTAINS "Life"
RETURN m

What type of index is generally best for this type of query?

  • ✓ TEXT index

  • ❏ LOOKUP index

  • ❏ RANGE index

  • ❏ Full-text index

Hint

This type of index is implemented for string property values.

Solution

The correct answer is TEXT index

Summary

In this lesson, you learned about the types of indexes that Neo4j supports. In the next lesson, you will learn how to create an index on a node property.