Databases : understand how to organize and to use data

Graph-oriented databases : Understanding

• Main difference between Relational and Graph : no tables
• Database is constituted of nodes (vertex) linked together thanks to edges (relationships)
• Nodes and edges can contain data
• Graph-oriented databases can be really close to the representation of a dependency graph drawn by a non computer-scientist

Modeling Graph databases

• What you need is : data and a meaningful way to connect them
• Example

Modeling Graph databases(2)

• Each node can contain data, formatted by a json data structure (not necessarily strictly enforced)
• Each relationship can also contain data using the same formatting

{
  "identity": 18,
  "labels": [
    "Person"
  ],
  "properties": {
    "name": "Tom Cruise",
    "born": 1962
  }
}

{
  "identity": 49,
  "start": 18,
  "end": 39,
  "type": "ACTED_IN",
  "properties": {
    "roles": [
          "Jerry Maguire"
        ]
  }}

Setting up Neo4j with docker-compose

version: '3.5'
services:
    # ...
    neo4j:
      image: neo4j:4.1
      restart: unless-stopped
      ports:
        - 10474:7474
        - 10687:7687
      volumes:
        - ./neo4j/conf:/conf
        - ./neo4j/data:/data
        - ./neo4j/import:/import
        - ./neo4j/logs:/logs
        - ./neo4j/plugins:/plugins
      environment:
        - NEO4J_dbms_memory_pagecache_size=1G
        - NEO4J_dbms.memory.heap.initial_size=1G
        - NEO4J_dbms_memory_heap_max__size=1G

Setting up Neo4j with docker-compose (2)

firewall-cmd --add-port=10474/tcp --permanent
firewall-cmd --add-port=10687/tcp --permanent
firewall-cmd --reload
docker-compose up -d neo4j
    

open this url in your browser :

Using Neo4J + Cypher

Graph Queries Queries : Create Node & Relationship

Example: create movie & actor nodes and relationship

CREATE (TheMatrix:Movie {title:'The Matrix', released:1999, tagline:'Welcome to the Real World'})
CREATE (Keanu:Person {name:'Keanu Reeves', born:1964})
CREATE (Keanu)-[:ACTED_IN {roles:['Neo']}]->(TheMatrix)

Graph Queries Queries : Create Node & Relationship

CREATE (TheMatrix:Movie {title:'The Matrix', released:1999, tagline:'Welcome to the Real World'})

• each node and relationship can be described thanks to a json ({title:'The Matrix'...} document, and a label (:Movie)

Graph Queries Queries : Browsing nodes and relationships

Example: search nodes with MATCH & RETURN

MATCH (m:Movie {title:'The Matrix'})
RETURN m

Exercise : complete "The Matrix" Casting

following the previous examples, input the following actors as nodes in the graph, and declare their relationships

• Carrie-Anne Moss {name:'Carrie-Anne Moss', born:1967}
• Laurence Fishburne {name:'Laurence Fishburne', born:1961}
• Hugo Weaving {name:'Hugo Weaving', born:1960}

declare that they all have acted in "The Matrix" Movie

Larger example : import a movie database

• Download the sample at this link
• Display all the database under the form of a graph
• Find 'Tom Cruise' and all the films he played in
• (To this result: ) add the actors linked to 'Tom Cruise' because they played a role in the same film

Connecting to neo4j from programming languages