Graph Database

Graph Database

1. Definition

• A database for storing and querying data in a data structure like graphs Composition

• Node

  • Specific entities, such as the movie “Harry Potter 2“,
• Relation
  • The connection between entities.
• Label
  • Abstract concepts that group similar nodes together, such as movie characters.
• Property
  • Specific information about the node or relationship, such as name, age, etc.

2. Neo4j

2.1 Introduction

• 知识图谱存储工具：RDF4J、Jena、Neo4j

• Neo4j is a robust, scalable, high-performance open source graph database

2.2 Neo4j-Important element

2.2.1 Structural Node

• ID: unique identifier
• Label: a form of pattern syntax to group nodes together
• Map(properties):
  • key: property name
  • value: property value

2.2.2 Properties

• Key: property name
• Value: property value

Value Type

• Boolean
• Byte
• Short
• Int
• Long
• Float
• Double
• Char
• String
• Array

2.2.3 Structural Relation

• ID
• Type
• Map(properties)
• ID of the start node
• ID of the end node

2.2.4 Path

• An sequence of nodes and relationships
• Composition
  • at least one node
  • connected relationships
• Often as a result of a query or traversal

• A path from nodeA to nodeB

3. Cypher

3.1 Introduction

• Cypher is a declarative graph query language that allows for expressive and efficient querying, updating and administering of the graph.

• Note： Keywords are not case sensitive

3.2 Cypher-Create

3.2.1 Create a node

1
2
3
4
5
6

CREATE(
	node:Movie
	{
		title:"The American President"
	}
)

3.2.2 Create relationships

1
2
3
4

CREATE
(<node1-name>:<node1-label-name>{<define-properties-list>})
-[<relationship-name>:<relationship-label-name>{<define-properties-list>}]
->(<node2-name>:<node2-label-name>{<define-properties-list>})

Exercise

• Create the relationship using cypher.

1
2
3
4

CREATE
	(n1:Person {name:'Oliver Stone'})
	-[r:DIRECTED]
	->(N2:Movie{title:"Wall Street"})

3.3 Cypher-MATCH

• Match

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

MATCH
    (<node-name>:<label-name>
    {
    <Property1-name>:<Property1-Value>
    ........
    <Propertyn-name>:<Propertyn-Value>
    })
    -[<relationship-name>:<relationship-label-name>{<define-properties-list>}]
    ->(<node-name>:<label-name>
    {
    <Property1-name>:<Property1-Value>
    ........
    <Propertyn-name>:<Propertyn-Value>
    })
RETURN ........

• Return
  • Variable, like node-name: node
  • Specific value, like node-name.property: node.name

3.3.1 Match node

• 查询所有节点

1
2

MATCH (n) 
RETURN n

• 查询所有电影title

1
2

MATCH (movie:Movie)
RETURN movie.title

3.3.2 Match related nodes

• 名叫Oliver Stone相关联的事物标题

1
2

MATCH (director {name: 'Oliver Stone'})--(movie) 
RETURN movie.title

• — : connected relationship of unknown relationship type and direction

• 与人物Oliver Stone相关联的电影标题

1
2

MATCH (:Person {name: 'Oliver Stone'})--(movie:Movie) 
RETURN movie.title

• 与人物Oliver Stone相关联的关系类型

1
2

MATCH (:Person {name: 'Oliver Stone'})-[r]->(movie) 
RETURN type(r)

• -> / <- : directed relation, > towards the tail entity

• 出演名为‘Wall Street’电影的演员姓名

1
2

MATCH (wallstreet:Movie {title: 'Wall Street'})<-[:ACTED_IN]-(actor) 
RETURN actor.name

3.3.4 Match on multiple relationship types

• 参演或导演名为Wall Street的电影的所有人物

1
2

MATCH (wallstreet:Movie {title: 'Wall Street'})<-[:ACTED_IN|:DIRECTED]-(person:Person) 
RETURN person.name

3.3.5 Match on relationship type and use a variable

• 名为Wall Street的电影的中所有的角色

1
2
3

MATCH (wallstreet:Movie {title: 'Wall Street'})<-
[r:ACTED_IN]-(actor) 
RETURN r.role

3.4 Match and Create

• when nodes already exist, we can use MATCH first, then CREATE

1
2
3
4
5

MATCH 
	(charlie:Person {name: 'Charlie Sheen'}), 
    (rob:Person {name: 'Rob Reiner'}) 
CREATE 
    (rob)-[:TYPE INCLUDING A SPACE]->(charlie)

Exercise

• 查询参演名为‘The American President’的所有演员姓名

1
2
3
4

MATCH
	(person:Person)-[:ACTED_IN]
	->(movie:Movie{title:"The American President"})
RETURN person.name

• 查询包含角色’Card Fox’的电影名

1
2
3
4

MATCH
	(actor)-[:ACTED{role:"Card Fox"}]
	->(movie:Movie)
RETURN movie.title

3.5 Cypher-DELETE

• Delete single node
• Delete all nodes and relationships
• Delete a node with all its relationships
• Delete relationships only

3.5.1 Delete single node

• 删除名为UNKNOWN的节点

1

MATCH (n:Person {name: 'UNKNOWN’}) DELETE n

3.5.2 Delete all nodes and relationships

• 删除数据库中所有节点及与其相连的关系

1

MATCH (n) DETACH DELETE n

3.5.3 Delete a node with all its relationships

• 删除名为Andy的节点和与其相连的所有关系

1

MATCH (n {name: 'Andy’}) DETACH DELETE n

3.5.4 Delete relationships only

• 删除Andy的所有KNOWS关系

1

MATCH (n {name: 'Andy'})-[r:KNOWS]->() DELETE r

Exercise

• 删除人物年龄为34岁的所有关系

1
2

MATCH (person:Person {age:34})-[r]->()
DELETE r

3.6 Cypher-UPDATE :

3.6.1 SET

• SET can be used with a map — provided as a literal, a parameter, or a node or relationship — to set properties.

• Update a property

1
2
3

MATCH (n {name: 'Andy'}) 
SET n.age = toString(n.age) 
RETURN n.name, n.age

3.6.2 multiple properties using one SET clause

1
2

MATCH (n {name: 'Andy'}) 
SET n.position = 'Developer', n.surname = 'Taylor'

3.6.3 Replace all properties using a map

1
2
3
4

MATCH (p {name: 'Peter'}) 
SET p = {name: 'Peter Smith’, 
	position: 'Entrepreneur'} 
RETURN p.name, p.age, p.position

Exercise

• 更新George的年龄为28

1
2
3

MATCH (p {name:'George'})
SET p = {name: 'George'
	age: 28}

Exercise

• Write cypher to get the second graph

1
2
3
4
5
6
7
8

MATCH 
	(actor1:Actor {name: 'Anthony Hopkins'}),
	(actor2:Actor {name: 'Hitchcock'}),
	(movie:Movie {title: 'Hitechcock'})
CREATE
	(actor1)-[r:ACTS_IN]
	->(movie)
DELETE actor2

• Query all relationship types connected by actor Anthony Hopkins

1
2
3

MATCH
	(actor:Actor {name: 'Anthony Hopkins'})-[r]-()
RETURN type(r)