Get Started With Gremlin

What's Gremlin?

Gremlin is used as the graph traversal language on Apache TinkerPop, which is a graph computing framework.

Basic Concepts

• Graph:

  Collection of Vertex and Edge.

• Element:

  Collection of Property.

• Vertex:

  Inherits from Element. Vertex is generally used to store the Property of entities.

• Edge:

  Inherits from Element. Edge is generally used to store the Property of relationship.

• Property

  A k-v pair, where the key is just like the column name in relational database, and the value is stored data in the Edge or Vertex.

  Note that the key must be a string.

• Label

  The class for Vertex or Edge. Vertexes and Edges having the same label means they have the same properties. Label can be regarded as the table name in a workplace in relational database.

Usage of Gremlin

Graph, Edge and Vertex

All query must start with a graph.

1
2
3
4
5
6
7
8

// g represents the whole graph

// Get all edges
g.E()
// Get all vertexes
g.V()
// Get 100 vertexes
g.V().limit(100)

Property

1
2
3
4
5
6
7

// Get all properties' names for all vertexes
g.V().properties().key()

// Get property value of all vertexes with property called "time"
g.V().properties('time').value()
// Same, but simpler
g.V().values('time')

ID

Get vertex or edge by id, or get id

1
2
3
4

// Get vertexes using id
g.V('id1')
// Get id for all vertexes
g.V().id() 

Label

Get vertex or edge by label

1
2
3
4

// Get vertexes which have label "software"
g.V().hasLabel('software')
// Get out-edge with label "develops" of vertex with id "46"
g.V('46').outE('develops')

Filter

In Gremlin, we can use hasXXX to filter specific edges or vertexes. Gremlin provides the following commands:

1. has(key, value): filter by property's key and value

2. has(label, key, value): filter by label, and property's key and value

3. has(key, expr): filter by label and an expr. For example, using g.V().has('age', gt(20)) we can get vertexes which have age property and whose value of age is larger than 20.

4. hasLabel(label1, label2, ...): filter by label. Elements with label in any one of the label list will pass the filter.

5. hasId(id1, id2, ...): filter by id, similar with hasLabel

6. hasKey(key1, key2, ...) and has(key): filter by key