Graph Database

2012-07-24 by Josh Adell (Meetup)

"Data are from Mars, RDBMs are from Venus."

Non-relational Databases

CouchDB
Cassandra
redis
MondoDB
riak

Graph Databases

OrientDB
Neo4J
Flock (Twitter)
others...

Graphs are everywhere:

  • street maps
  • computer network
  • social network
  • relational databases
    • records == nodes
    • foreign keys == edges
    • ORMs translate to in-memory graph
  • whiteboard friendly
  • natural language processing
    • nouns == nodes
    • verbs == relationships
  • access control list
  • LinkIn uses Neo4J to find relations on the fly
  • Twitter uses custom-built graph database
  • FrostyMug beer recommendation engine

Neo4J

  • RESTful interface
  • ACID complient
  • High availability
  • Scalable
  • AGPL licenced
  • http://console.neo4j.org/
  • Lucene indexing engine by default
  • Each node is an arbitrary bag of properties
    • Edges are special nodes with two properties (left,right)
    • Not really types that Neo4J, need to be managed by application
    • Which index a node is in can determine it's "type"
  • Jeff is a graph serialization format
  • Cypher is a query language proposal by Neo4J

Cypher Queries Examples

# http://tinyurl.com/c65d99w
# Kevin Bacon
START s=node:actors(name="Keanu Reeves"),
e=node:actors(name="Kevin Bacon")
MATCH p = shortestPath(s-[*]-e)
RETURN p, length(p)
# http://tinyurl.com/cyn3rkx
# no need to worry about nested sub-groups

# start with user 3
START u=node:users(name="User 3")

# find all groups the user(s) is in
MATCH u-[:belongs_to*]->g

# return group(s) that user 3 is in
RETURN g
# http://tinyurl.com/dx7onro
START u=node:users(name="User 3"),
o=node:objects(name="Home")
# find zero or more belongs_to relationships
# that can_read the home page
# for user 1, g is u (zero or more...)
MATCH u-[:belongs_to*0..]->g,
g-[:can_read]->o
RETURN g
# http://tinyurl.com/bwtyhvt
START u=node:users(name="User 3"),
o=node:objects(name="Home")
MATCH u-[:belongs_to*0..]->g,
g-[:can_read]->o,
# optional denied relationship
u-[d?:denied*]->o
WHERE d is null
RETURN g

Resturant Example

Companies have brands, locations, location groups
Brands have locations, location groups
Location groups have locations

# http://tinyurl.com/cxm4heh
START c=node:companies(name="Company 1")
MATCH c-[:HAS*]->1
WHERE l.type = 'location'
RETURN l
ORDER BY l.name
# http://tinyurl.com/cl537w6
START b=node(3)
MATCH b<-[:HAS*]-c-[:HAS*]->l<-[h?:HAS*]-b
WHERE h IS NULL AND l.type='location'
RETURN l
ORDER BY l.name