Graph Databases

Immanuel Trummer

[email protected]

www.itrummer.org
 Outlook: 

Beyond Relational Data

• Graph data

• Data streams

• Spatial data

Slides by Immanuel Trummer, Cornell University

 Outlook: 

Beyond Relational Data

• Graph data

• Data streams

• Spatial data

Slides by Immanuel Trummer, Cornell University

 Reading List

• "Graph Databases" by I. Robinson et al.

• "Graph Databases Comparison: AllegroGraph, ArangoDB,

InfiniteGraph, Neo4J, and OrientDB" 

by Fernandes and Bernardino.

• http://www.Neo4j.com

Slides by Immanuel Trummer, Cornell University

 Graph Data

• A set of nodes and a set of edges connecting nodes

• Nodes and edges can be associated with labels

• Nodes and edges can be associated with properties

Slides by Immanuel Trummer, Cornell University

 Example (Toy) Graph
Name: "Marc"

:Student
Name: "CS4320" Name: "Cornell"
:Enrolled

:Course :University
:Offers

:Enrolled

:Student

Name: "Maria"
Slides by Immanuel Trummer, Cornell University
 Motivation

• Social networks

• Knowledge graphs

• Communication graphs

• Road networks

• ...

Slides by Immanuel Trummer, Cornell University

Communication Structure of the Internet as Graph.
Visualization of Facebook Connections as Graph.
NYC Metro Graph
NYC Metro Graph

All Paths?
 How to represent 

Graph as Relational DB?

Slides by Immanuel Trummer, Cornell University

Relational Representation

• CREATE TABLE Stations(


StationID int primary key, name text);

• CREATE TABLE Connected(


StationID1 int, StationID2, 

primary key (StationID1, StationID2),

foreign key (StationID1) references Stations(StationID1),

foreign key (StationID2) references Stations(StationID2)

);

Slides by Immanuel Trummer, Cornell University

Query: Find Paths from
Port Authority to NYU?

Slides by Immanuel Trummer, Cornell University

Find Paths from P to N
• SELECT * from Connected C1 

join Connected C2 on (C1.stationid2 = C2.stationid1)

join Connected C3 on (C2.stationid2 = C3.stationid1)

... join Connected Cn ...

WHERE C1.name = 'Port Authority' 

and Cn.name = 'NYU'

• Retrieves paths with a fixed length (can iterate)

• (Can be solved better with advanced SQL features)

Slides by Immanuel Trummer, Cornell University

(Intermediate) Conclusions

• Storing graph data in relational DBMS is possible

• But querying graphs via vanilla SQL is inconvenient

• Also, may increase efficiency by graph specialization

Slides by Immanuel Trummer, Cornell University

Graph Database Systems

Slides by Immanuel Trummer, Cornell University

Graph Database Systems

Slides by Immanuel Trummer, Cornell University

 Cypher

• Graph query language used by Neo4j

• Allows creating/updating nodes and relationships

• Allows searching graphs for complex patterns

• Aggregation, filtering, sub-queries etc.

• Inspired by SQL in some aspects

Slides by Immanuel Trummer, Cornell University

 Creating Nodes

• CREATE ()

Create node without labels or properties

• CREATE (:Student)

Create node labeled as student, no properties

• CREATE (:Student {name : 'Marc'})


Create node labeled as student, name set to 'Marc'

Slides by Immanuel Trummer, Cornell University

 Finding Nodes

• MATCH (m:Student {name : 'Marc'})

• Finds nodes labeled as "Student"

• Name property must be set to "Marc"

• Match result is assigned to variable m

• Variable m can be used in remaining query

Slides by Immanuel Trummer, Cornell University

Creating Relationships
• MATCH (a:Student {name: 'Marc'}), 

(b:Course {name: 'CS4320'})

CREATE (a)-[:Enrolled {semester: 'FS20'}]->(b)

• Matches a to students with name "Marc"

• Matches b to courses with name "CS4320"

• Inserts edge from a to b with label "Enrolled"

• Edge has property "semester" set to "FS20"

Slides by Immanuel Trummer, Cornell University

 Updating Nodes

• MATCH (m:Student {name: 'Marc'})


SET m:Alumnus

• Changes label of Marc from Student to Alumnus

• MATCH (m:Student {name: 'Marc'})


SET m.name = 'Marcus'

• Changes value of name property to "Marcus"

Slides by Immanuel Trummer, Cornell University

 Finding Relationships
• MATCH (a:Student {name: 'Marc'})

-[e:Enrolled {semester: 'FS20'}]-

(b:Course {name: 'CS4320'})

• Find edges connecting nodes a and b such that

• Node a is a student with name 'Marc'

• Node b is a course with name 'CS4320'

• Edge labeled "Enrolled", property semester is "FS20"

• Assign resulting edges to variable e

Slides by Immanuel Trummer, Cornell University

Updating Relationships

• MATCH (a:Student {name: 'Marc'})


-[e:Enrolled {semester: 'FS20'}]-

(b:Course {name: 'CS4320'})

SET e.semester = 'FS21'

• Get edge representing enrollment of Marc in CS4320

• Update value of semester property to "FS21"

Slides by Immanuel Trummer, Cornell University

 Deletions

• MATCH (a:Student {name: 'Marc'})


DELETE a

• Deletes students with name "Marc" from the database

Slides by Immanuel Trummer, Cornell University

(Demo)

Slides by Immanuel Trummer, Cornell University

Exercise: Create Graph DB

• Create a graph DB representing the following situation

• Ithaca and Binghamton are cities located in NY state

• Cornell University is located in Ithaca

• Cornell offers a course with name "CS4320"

Slides by Immanuel Trummer, Cornell University

Pattern-Based Retrieval

• MATCH ( :Student {name: 'Marc'} )


-[:friendsWith]-> (s:Student)

RETURN s

• Returns all friends (students) of student Marc


Slides by Immanuel Trummer, Cornell University

Pattern-Based Retrieval

• MATCH ( :Student {name: 'Marc'} )


-[:friendsWith*]-> (s:Student)

RETURN s

• Returns all friends (students) of Marc, their friends, the

friends of their friends, etc.

Slides by Immanuel Trummer, Cornell University

Pattern-Based Retrieval

• MATCH ( :Student {name: 'Marc'} )


-[:friendsWith*0..2]-> (s:Student)

RETURN s

• Any suggestions: what does this (probably) return?


Slides by Immanuel Trummer, Cornell University

 Aggregation

• MATCH ( :Student {name: 'Marc'} )


-[:friendsWith]-> (:Student)

RETURN count(*)

• Count number of friends of Marc

Slides by Immanuel Trummer, Cornell University

 Complex Patterns

• MATCH (s1:Student) -[:friendsWith]->(s2:Student),


(s1)-[:Enrolled]->(c:Course), (s2)-[:Enrolled]->(c)

WHERE s1.name IN ['Marc', 'Maria'] 

AND NOT c.name = 'CS4320'

RETURN s2

• Friends of Marc and Maria who have at least one

course in common with them, excluding CS4320

Slides by Immanuel Trummer, Cornell University

Retrieve Courses Taken
by At Least One Student
Who Also Takes CS4320!

Slides by Immanuel Trummer, Cornell University

 (Initial Example)

• MATCH (s1:Station) -[:Connected*]- (s2:Station)


RETURN *

Slides by Immanuel Trummer, Cornell University

 Data Layout

• In-memory data layout is optimized for fast traversals

• Nodes stored with label, properties, and edge references

• Node stores list of incoming and outgoing edges

• Edges stored with label, properties, and node references

Slides by Immanuel Trummer, Cornell University

 Query Processing

• Query plans composed from standard operators

• Most known from SQL: filtering, projection, ...

• A few graph-specific operators (e.g., shortest path)

• Can use indices to retrieve specific nodes/edges

• Query plans are selected via cost-based optimization

Slides by Immanuel Trummer, Cornell University

Transaction Processing

• Neo4j supports read-committed isolation by default

• Acquire locks manually to achieve higher isolation level

• Uses logging to persistent storage to achieve durability

• Overall: can support ACID

Slides by Immanuel Trummer, Cornell University