HugeGraph Gremlin

HugeGraph支持Apache TinkerPop3的图形遍历查询语言Gremlin。 SQL是关系型数据库查询语言，而Gremlin是一种通用的图数据库查询语言，Gremlin可用于创建图的实体（Vertex和Edge）、修改实体内部属性、删除实体，也可执行图的查询操作。

Gremlin可用于创建图的实体（Vertex和Edge）、修改实体内部属性、删除实体，更主要的是可用于执行图的查询及分析操作。

TinkerPop Features

HugeGraph实现了TinkerPop框架，但是并没有实现TinkerPop所有的特性。

下表列出HugeGraph对TinkerPop各种特性的支持情况：

Graph Features

Name Description Support
Computer Determines if the {@code Graph} implementation supports {@link GraphComputer} based processing false
Transactions Determines if the {@code Graph} implementations supports transactions. true
Persistence Determines if the {@code Graph} implementation supports persisting it’s contents natively to disk.This feature does not refer to every graph’s ability to write to disk via the Gremlin IO packages(.e.g. GraphML), unless the graph natively persists to disk via those options somehow. For example,TinkerGraph does not support this feature as it is a pure in-sideEffects graph. true
ThreadedTransactions Determines if the {@code Graph} implementation supports threaded transactions which allow a transactionto be executed across multiple threads via {@link Transaction#createThreadedTx()}. false
ConcurrentAccess Determines if the {@code Graph} implementation supports more than one connection to the same instance at the same time. For example, Neo4j embedded does not support this feature because concurrent access to the same database files by multiple instances is not possible. However, Neo4j HA could support this feature as each new {@code Graph} instance coordinates with the Neo4j cluster allowing multiple instances to operate on the same database. false

Vertex Features

Name Description Support
UserSuppliedIds Determines if an {@link Element} can have a user defined identifier. Implementation that do not support this feature will be expected to auto-generate unique identifiers. In other words, if the {@link Graph} allows {@code graph.addVertex(id,x)} to work and thus set the identifier of the newly added {@link Vertex} to the value of {@code x} then this feature should return true. In this case, {@code x} is assumed to be an identifier datat ype that the {@link Graph} will accept. false
NumericIds Determines if an {@link Element} has numeric identifiers as their internal representation. In other words,if the value returned from {@link Element#id()} is a numeric value then this method should be return {@code true}. Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. false
StringIds Determines if an {@link Element} has string identifiers as their internal representation. In other words, if the value returned from {@link Element#id()} is a string value then this method should be return {@code true}. Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. false
UuidIds Determines if an {@link Element} has UUID identifiers as their internal representation. In other words,if the value returned from {@link Element#id()} is a {@link UUID} value then this method should be return {@code true}.Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. false
CustomIds Determines if an {@link Element} has a specific custom object as their internal representation.In other words, if the value returned from {@link Element#id()} is a type defined by the graph implementations, such as OrientDB’s {@code Rid}, then this method should be return {@code true}.Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. false
AnyIds Determines if an {@link Element} any Java object is a suitable identifier. TinkerGraph is a good example of a {@link Graph} that can support this feature, as it can use any {@link Object} as a value for the identifier. Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. This setting should only return {@code true} if {@link #supportsUserSuppliedIds()} is {@code true}. false
AddProperty Determines if an {@link Element} allows properties to be added. This feature is set independently from supporting “data types” and refers to support of calls to {@link Element#property(String, Object)}. true
RemoveProperty Determines if an {@link Element} allows properties to be removed. true
AddVertices Determines if a {@link Vertex} can be added to the {@code Graph}. true
MultiProperties Determines if a {@link Vertex} can support multiple properties with the same key. false
DuplicateMultiProperties Determines if a {@link Vertex} can support non-unique values on the same key. For this valueto be {@code true}, then {@link #supportsMetaProperties()} must also return true. By default this method, just returns what {@link #supportsMultiProperties()} returns. false
MetaProperties Determines if a {@link Vertex} can support properties on vertex properties. It is assumed that a graph will support all the same data types for meta-properties that are supported for regular properties. false
RemoveVertices Determines if a {@link Vertex} can be removed from the {@code Graph}. true

Edge Features

Name Description Support
UserSuppliedIds Determines if an {@link Element} can have a user defined identifier. Implementation that do not support this feature will be expected to auto-generate unique identifiers. In other words, if the {@link Graph} allows {@code graph.addVertex(id,x)} to work and thus set the identifier of the newly added {@link Vertex} to the value of {@code x} then this feature should return true. In this case, {@code x} is assumed to be an identifier datat ype that the {@link Graph} will accept. false
NumericIds Determines if an {@link Element} has numeric identifiers as their internal representation. In other words,if the value returned from {@link Element#id()} is a numeric value then this method should be return {@code true}. Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. false
StringIds Determines if an {@link Element} has string identifiers as their internal representation. In other words, if the value returned from {@link Element#id()} is a string value then this method should be return {@code true}. Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. false
UuidIds Determines if an {@link Element} has UUID identifiers as their internal representation. In other words,if the value returned from {@link Element#id()} is a {@link UUID} value then this method should be return {@code true}.Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. false
CustomIds Determines if an {@link Element} has a specific custom object as their internal representation.In other words, if the value returned from {@link Element#id()} is a type defined by the graph implementations, such as OrientDB’s {@code Rid}, then this method should be return {@code true}.Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. false
AnyIds Determines if an {@link Element} any Java object is a suitable identifier. TinkerGraph is a good example of a {@link Graph} that can support this feature, as it can use any {@link Object} as a value for the identifier. Note that this feature is most generally used for determining the appropriate tests to execute in the Gremlin Test Suite. This setting should only return {@code true} if {@link #supportsUserSuppliedIds()} is {@code true}. false
AddProperty Determines if an {@link Element} allows properties to be added. This feature is set independently from supporting “data types” and refers to support of calls to {@link Element#property(String, Object)}. true
RemoveProperty Determines if an {@link Element} allows properties to be removed. true
AddEdges Determines if an {@link Edge} can be added to a {@code Vertex}. true
RemoveEdges Determines if an {@link Edge} can be removed from a {@code Vertex}. true

Data Type Features

Name Description Support
BooleanValues true
ByteValues true
DoubleValues true
FloatValues true
IntegerValues true
LongValues true
MapValues Supports setting of a {@code Map} value. The assumption is that the {@code Map} can containarbitrary serializable values that may or may not be defined as a feature itself false
MixedListValues Supports setting of a {@code List} value. The assumption is that the {@code List} can containarbitrary serializable values that may or may not be defined as a feature itself. As this{@code List} is “mixed” it does not need to contain objects of the same type. false
BooleanArrayValues false
ByteArrayValues true
DoubleArrayValues false
FloatArrayValues false
IntegerArrayValues false
LongArrayValues false
SerializableValues false
StringArrayValues false
StringValues true
UniformListValues Supports setting of a {@code List} value. The assumption is that the {@code List} can containarbitrary serializable values that may or may not be defined as a feature itself. As this{@code List} is “uniform” it must contain objects of the same type. false

Gremlin的步骤

HugeGraph支持Gremlin的所有步骤。有关Gremlin的完整参考信息，请参与Gremlin官网。

步骤 说明 文档
addE 在两个顶点之间添加边 addE step
addV 将顶点添加到图形 addV step
and 确保所有遍历都返回值 and step
as 用于向步骤的输出分配变量的步骤调制器 as step
by 与group和order配合使用的步骤调制器 by step
coalesce 返回第一个返回结果的遍历 coalesce step
constant 返回常量值。 与coalesce配合使用 constant step
count 从遍历返回计数 count step
dedup 返回已删除重复内容的值 dedup step
drop 丢弃值（顶点/边缘） drop step
fold 充当用于计算结果聚合值的屏障 fold step
group 根据指定的标签将值分组 group step
has 用于筛选属性、顶点和边缘。 支持hasLabel、hasId、hasNot 和 has 变体 has step
inject 将值注入流中 inject step
is 用于通过布尔表达式执行筛选器 is step
limit 用于限制遍历中的项数 limit step
local 本地包装遍历的某个部分，类似于子查询 local step
not 用于生成筛选器的求反结果 not step
optional 如果生成了某个结果，则返回指定遍历的结果，否则返回调用元素 optional step
or 确保至少有一个遍历会返回值 or step
order 按指定的排序顺序返回结果 order step
path 返回遍历的完整路径 path step
project 将属性投影为映射 project step
properties 返回指定标签的属性 properties step
range 根据指定的值范围进行筛选 range step
repeat 将步骤重复指定的次数。 用于循环 repeat step
sample 用于对遍历返回的结果采样 sample step
select 用于投影遍历返回的结果 select step
store 用于遍历返回的非阻塞聚合 store step
tree 将顶点中的路径聚合到树中 tree step
unfold 将迭代器作为步骤展开 unfold step
union 合并多个遍历返回的结果 union step
V 包括顶点与边之间的遍历所需的步骤：V、E、out、in、both、outE、inE、bothE、outV、inV、bothV 和 otherV order step
where 用于筛选遍历返回的结果。 支持 eq、neq、lt、lte、gt、gte 和 between 运算符 where step