4 Rubyist

Graph Database

why?

Similar things

------------------------------------------------------

*| |*

Students * ----- * Courses 1 ------ * Achievements

Students 1 ---- * UserCourses * ---- 1 Courses

Students 1 ---- * UserAchievements * ---- 1 UserAchievements

what?

In short

Data stored as nodes and relationships
Index-free adjacency
Examples:
TAO (Facebook),
Neo4j (Open-sourced),
Graphd (Freebase), etc

Neo4j

A graph database

http://www.neo4j.org/

Other interesting things

REST API for db operations
Graph query language: Cypher
Visualizer

try!

Scenario

Person 1 --- * PersonFriend * --- 1 Person

Create nodes & relationships

Nodes

CREATE (person {name : 'John'})RETURN person;

Relationship

START a=node(9), b=node(10)CREATE a-[rel:friend]->bRETURN rel

Query

Some simple one :)

START n=node(*)
RETURN n

START n=node:nodes(name = "John")
RETURN n

Visualize

Live

http://www.neo4j.org/learn/try

why_again?

Expressiveness

Who'are John's friends

SELECT Person.name as PERSON, PersonFriend.friend_id as FRIENDFROM PersonINNER JOIN PersonFriendON Person.id = PersonFriend.person_idWHERE Person.name =  'John';

Cyper

START person=node:Person(name = 'John')MATCH person-[:friend]->friendRETURN friend

Expressiveness

John's friends of friends

SELECT Person.name as PERSON, pf1.friend_id as COMMON,  pf2.friend_id as FoFFROM Person
INNER JOIN PersonFriend pf1 ON Person.id = pf1.person_idINNER JOIN PersonFriend pf2 ON pf1.friend_id = pf2.person_idWHERE Person.name =  'John';

Cyper

START person=node:Person(name = 'John')MATCH person-[:friend]->common_friend-[:friend]->friendRETURN friend, common_friend

Performance

Find friends-of-friends in a social network,
to a maximum depth of five
Social network containing 1,000,000 people,
each with approximately 50 friends
Result: