Agenda
- Use cases for graph databases
- Advantages of graph databases
- Which graph database is good for me?
- Play time with Neo4j!
- Play time with Cayley!
Agenda
- Use cases for graph databases
- Advantages of graph databases
- Which graph database is good for me?
- Play time with Neo4j!
- Play time with Cayley!
Ubuntu Package Repository
Agenda
- Use cases for graph databases
- Advantages of graph databases
- Which graph database is good for me?
- Play time with Neo4j!
- Play time with Cayley!
Performance
Finding extended friends in a relational database vs graph database.
| Depth |
RDBMS execution time |
Graph execution time |
Records returned |
| 2 |
0.01 sec |
0.01 sec |
2500 |
| 3 |
30 sec |
0.16 sec |
110,000 |
| 4 |
1543 sec |
1.3 sec |
600,000 |
| 5 |
Unfinished |
2.1 sec |
800,000 |
Flexiblity
Reduce the cost of change
Agenda
- Use cases for graph databases
- Advantages of graph databases
- Which graph database is good for me?
- Play time with Neo4j!
- Play time with Cayley!
What is a Graph database?
"a graph database is a database that uses graph structures for semantic queries with nodes, edges and properties to represent and store data."
Comparison - Neo4j
| First commit |
2002 |
| Language |
Java |
| Type |
Property graph |
| Pluggable backend |
No |
| Query language |
Cypher and Gremlin |
| License and price |
Community edition - GPL 3, free |
|
Enterprise edition - AGPL 3, 20k a year |
| Created by |
Neo4j, Swedish company. HQ in Silicon Valley |
| Community |
Large - books, events, training, IRC, Slack |
| Time to first hello world |
10 minutes |
Comparison - Titan
| First commit |
2012 |
| Language |
Java |
| Type |
Property graph |
| Pluggable backend |
Yes - Cassandra, HBase, BerkeleyDB, Amazon DynamoDB |
| Query language |
Gremlin |
| License and price |
Apache 2, free |
| Created by |
Aurelius, startup which got acquired by DataStax. DataStax's HQ is in San Francisco |
| Community |
Medium - Google group |
| Time to first hello world |
a few hours |
Comparison - Cayley
| First commit |
2014 |
| Language |
Go |
| Type |
RDF (triple) store |
| Pluggable backend |
Yes - LevelDB, Bolt, PostgreSQL, MongoDB, in-memory |
| Query language |
Javascript, with a Gremlin-inspired graph object |
| License and price |
Apache 2, free |
| Created by |
Ex-Googlers that were part of Freebase and Google Knowledge Graph |
| Community |
Small - Google group, IRC, Slack |
| Time to first hello world |
1 minute |
Agenda
- Use cases for graph databases
- Advantages of graph databases
- Which graph database is good for me?
- Play time with Neo4j!
- Play time with Cayley!
Agenda
- Use cases for graph databases
- Advantages of graph databases
- Which graph database is good for me?
- Play time with Neo4j!
- Play time with Cayley!
Quads
character:ice-king "has a crush on" character:princess-bubblegum .
character:ice-king "has a crush on" character:marceline-abadeer .
Subject Predicate Object
Find everyone that Ice King has a crush on
g.V("character:ice-king").Out("has a crush on").All()
{
"result": [ {
"id": "character:marceline-abadeer"
},
{
"id": "character:princess-bubblegum"
}
]
}
Find everyone that have a crush on Princess Bubblegum
g.V("character:princess-bubblegum").In("has a crush on").All()
{
"result": [ {
"id": "character:ice-king"
}
]
}
cayley load --config=cayley.cfg --quads=db.nq
Find all the haters
g.V().In("hates").All()
Find all the haters
g.V().In("hates").All()
{
"result": [
{
"id": "character:lumpy-space-princess",
},
{
"id": "character:bmo",
},
{
"id": "character:marceline-abadeer",
},
{
"id": "character:finn",
},
{
"id": "character:lady-rainicorn",
},
{
"id": "character:princess-bubblegum",
},
{
"id": "character:jake",
}
]
}
Find all the haters (and tag who they hate)
g.V().Tag("name").In("hates").All()
{
"result": [
{
"id": "character:lumpy-space-princess",
"name": "character:ice-king"
},
{
"id": "character:bmo",
"name": "character:ice-king"
},
{
"id": "character:marceline-abadeer",
"name": "character:ice-king"
},
{
"id": "character:finn",
"name": "character:ice-king"
},
{
"id": "character:lady-rainicorn",
"name": "character:ice-king"
},
{
"id": "character:princess-bubblegum",
"name": "character:ice-king"
},
{
"id": "character:jake",
"name": "character:ice-king"
}
]
}
Who hates Ice King?
g.V().Has("hates", "character:ice-king"").All()
{
"result": [
{
"id": "character:lumpy-space-princess"
},
{
"id": "character:bmo"
},
{
"id": "character:marceline-abadeer"
},
{
"id": "character:finn"
},
{
"id": "character:lady-rainicorn"
},
{
"id": "character:princess-bubblegum"
},
{
"id": "character:jake"
}
]
}
Find all haters that lives with BMO
Find all haters that lives with BMO
haters = g.V().In("hates")
lives_with_bmo = g.V("character:bmo").In("lives with")
haters.And(lives_with_bmo).All()
Find all haters that lives with BMO
haters = g.V().In("hates")
lives_with_bmo = g.V("character:bmo").In("lives with")
haters.And(lives_with_bmo).All()
{
"result": [
{
"id": "character:finn"
},
{
"id": "character:jake"
}
]
}
In the pipeline
- Typed Quad values
- Update nodes
- Aggregation
- Partial matching
Learn more about Cayley
- Website - https://cayley.io
- Forum - https://discourse.cayley.io
- IRC - #cayley on freenode
- Github - https://github.com/cayleygraph/cayley
Videos
- Intro to graph dbs and deep dive into Cayley's query language - Video
- Building a knowledge graph with Cayley - Video
