A walk in graph databases v1.0

A Walk in Graph Databases
Pierre De Wilde
4 May 2012
Global Brain Institute
VUB - ECCO Group

The Law of the Hammer

If the only tool you have is a hammer,
everything looks like a nail.
Abraham Maslow - The Psychology of Science - 1966

The Law of the Relational Database

If the only tool you have is a relational database,
everything looks like a table.
A Walk in Graph Databases - 2012

doesn't
One size fits all

Scalability issue
Scale up
Scale out

Index-intensive issue
Find data
Join data

NoSQL ?! No SQL ? Not only SQL !

Scalability solutions
Key-value stores
Column databases
Document databases

Index-intensive solution
Graph databases

Query language for relational databases

SQL
ISUD or CRUD

Traversal graph
Query language for relational databases

Gremlin is a graph traversal language

Map of walk
Warm up
Property Graph
Graph Database

Walk with Gremlin
Graph Manipulation
Graph Traversal

Stretching
Linked Data
Global Graph

Graph

G = (V, E)

--. .-. .- .--. ....

One graph doesn't fit all

Marko A. Rodriguez and Peter Neubauer - Constructions from Dots and Lines - 2010

Property graph

A property graph is a directed, labeled, attributed, multi graph.

Anatomy of a vertex

A vertex is composed of

- an unique identifier (id)
- a collection of properties
- a set of incoming edges (inE)
- a set of outgoing edges (outE)

Anatomy of an edge

An edge is composed of

- an unique identifier (id)
- an outgoing vertex (outV)
- a label
- an incoming vertex (inV)
- a collection of properties

Key feature of a graph database

Index-free adjacency

Some graph database vendors

Neo4j from Neo Technology
http://neo4j.org/

OrientDB from Orient Technologies
http://www.orientdb.org/

Dex from Sparsity-Technologies
http://www.sparsity-technologies.com/dex

InfiniteGraph from Objectivity, Inc.
http://www.infinitegraph.com/

TinkerPop

Open source project in the graph space

TinkerPop family

https://github.com/tinkerpop

Gremlin

$ gremlin.sh

,,,/
(o o)
-----oOOo-(_)-oOOo-----
gremlin>


Connect to a graph database

gremlin> g = new TinkerGraph(name)

gremlin> g = new Neo4jGraph(name)

gremlin> g = new OrientGraph(name)

gremlin> g = new DexGraph(name)

gremlin> g = new IGGraph(name)

Add a vertex / an edge

gremlin> v1 = g.addVertex()
gremlin> v2 = g.addVertex()
...

gremlin> g.addEdge(v1, 'knows', v2)
...

gremlin> g.loadGraphML(url)

Update a vertex

gremlin> v = g.getVertex(1)
==>v[1]

gremlin> v.getPropertyKeys()
==>age
==>name

gremlin> v.getProperty('name')
==>marko
gremlin> v.getProperty('age')
==>29
gremlin> v.setProperty('age',32)
==>32

gremlin> v.age
==>32
gremlin> v.name
==>marko

Update an edge

gremlin> e = g.getEdge(8)
==>e[8][1-knows->4]

gremlin> e.getPropertyKeys()
==>weight

gremlin> e.getProperty('weight')
==>1.0
gremlin> e.setProperty('weigth',0.9)
==>0.9

gremlin> e.map()
==>weigth=0.9
==>weight=1.0

gremlin> e.removeProperty('weigth')
==>0.9

Remove a vertex

gremlin> v = g.getVertex(3)
==>v[3]
gremlin> g.removeVertex(v)
==>null

Remove an edge

gremlin> e = g.getEdge(10)
==>e[10][4-created->5]
gremlin> g.removeEdge(e)
==>null

Disconnect from the graph database

gremlin> g.shutdown()

Graph traversal

Jump

- from vertex to edge
- from edge to vertex
- from vertex to vertex

Graph traversal: starting the traversal

gremlin> g.v(1)
==>v[1]

Graph traversal: outgoing edges

gremlin> g.v(1).outE
==>e[7][1-knows->2]
==>e[9][1-created->3]
==>e[8][1-knows->4]

Graph traversal: incoming vertices

gremlin> g.v(1).outE.inV
==>v[2]
==>v[4]
==>v[3]

Graph traversal: outgoing edges (cont.)

gremlin> g.v(1).outE.inV.outE
==>e[10][4-created->5]
==>e[11][4-created->3]

Graph traversal: incoming vertices (cont.)

gremlin> g.v(1).outE.inV.outE.inV
==>v[5]
==>v[3]

Graph traversal: ending the traversal

gremlin> g.v(1).outE.inV.outE.inV.outE

Graph traversal: starting vertex

gremlin> g.v(1)
==>v[1]

Graph traversal: adjacent vertices

gremlin> g.v(1).out
==>v[2]
==>v[4]
==>v[3]

Graph traversal: adjacent vertices (cont.)

gremlin> g.v(1).out.out
==>v[5]
==>v[3]

Graph traversal: labeled outgoing edges

gremlin> g.v(1).outE('created')
==>e[9][1-created->3]

Graph traversal: labeled adjacent vertices

gremlin> g.v(1).outE('created').inV
==>v[3]

gremlin> g.v(1).out('created')
==>v[3]

Graph traversal: labeled adjacent (cont.)

gremlin> g.v(1).out('created').in('created')
==>v[1]
==>v[4]
==>v[6]

Graph traversal and ...

index
transform
filter
compute
manipulate
loop
path

Graph traversal and index

gremlin> g.idx('vertices')[[name:'marko']]
==>v[1]

Graph traversal and transform

gremlin> g.v(1).outE.label.dedup
==>knows
==>created

gremlin> g.v(1).out('knows').name
==>vadas
==>josh

Graph traversal and filter

gremlin> g.v(1).out('knows').age
==>27
==>32

gremlin> g.v(1).out('knows').filter{it
.age>30}.age
==>32

Graph traversal and compute

gremlin> g.v(1).outE.weight
==>0.5
==>1.0
==>0.4

gremlin> g.v(1).outE.weight.mean()
==>0.6333333353201548

Graph traversal and manipulate

gremlin> g.v(1).outE.sideEffect{it.
weight+=0.1}.weight
==>0.6
==>1.1
==>0.5

Graph traversal and loop

gremlin> g.v(1).out.loop(1){it.loops<3}
==>v[5]
==>v[3]

Graph traversal and path

gremlin> g.v(1).outE.inV.path
==>[v[1], e[7][1-knows->2], v[2]]
==>[v[1], e[8][1-knows->4], v[4]]
==>[v[1], e[9][1-created->3], v[3]]

gremlin> g.v(1).out.path
==>[v[1], v[2]]
==>[v[1], v[4]]
==>[v[1], v[3]]

Global traversal: in-degree distribution

gremlin> m=[:].withDefault{0}; g.V.
sideEffect{m[it.in.count()]+=1}.
iterate(); m.sort()
==>0=2
==>1=3
==>3=1

Walk is ending

flexible

Linked Data

http://www.w3.org/DesignIssues/LinkedData.html

Linked Data cloud

Linking Open Data cloud diagram, by Richard Cyganiak and Anja Jentzsch. http://lod-cloud.net/

Linked Data initiative

http://freeyourmetadata.org/

Linked Data and Gremlin

+ +
gremlin> g = new SparqlRepositorySailGraph("http://dbpedia.org/sparql")

gremlin> v = g.v(' http://dbpedia.org/resource/Global_brain')
==>v[http://dbpedia.org/resource/Global_brain]

gremlin> v.out('http://www.w3.org/2000/01/rdf-schema#comment').has('lang',
'en').value
==>The Global Brain is a metaphor for the worldwide intelligent network...

gremlin> v.inE('http://dbpedia.org/ontology/knownFor').outV
==>v[http://dbpedia.org/resource/Francis_Heylighen]

gremlin> v.inE('http://dbpedia.org/ontology/knownFor').outV.outE('http://
dbpedia.org/ontology/knownFor').inV
==>v[http://dbpedia.org/resource/Self-organization]
==>v[http://dbpedia.org/resource/Memetics]
==>v[http://dbpedia.org/resource/Global_brain]

Global Graph

Internet => net of computers

Word Wide Web => web of documents

Giant Global Graph => graph of metadata

I called this graph the Semantic Web, but maybe it should have been Giant Global Graph.

Tim Berners-Lee - timbl's blog - 2007

http://tinkerpop.com

Logos created by Ketrina Yim for TinkerPop geeks
Images created by Flickr Creative Commons Artists
Graphs created by Memotive Concept Mapping tool

A walk in graph databases v1.0

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to A walk in graph databases v1.0 (20)

A walk in graph databases v1.0