Match

GeaFlow支持以下语法 ： * Match * Regex-Match * Return * Let * SubQuery * Continue-Match

Syntax

MatchStatement: MATCH PathPatthern (',' PathPatthern)* [WHERE boolExpr]

PathPatthern: Node ([Edge] Node)*
Node: '(' Identifier [ ':' StringLiteral ] [ WHERE boolExpr] ')'
Edge: '-' '[' Identifier [ ':' StringLiteral ] [ WHERE boolExpr] ']' '-'
         | '-' '[' Identifier [ ':' StringLiteral ] [ WHERE boolExpr] ']' '->'
         | '<-' '[' Identifier [ ':' StringLiteral ] [ WHERE boolExpr] ']' '-'

Node

匹配图上的点,可以指定点的类型以及对点的过滤条件。

Edge

匹配图上的边，类似Node节点可以指定边的类型以及对边的过滤条件。和Node不同的是,边需要指定方向，边的方向包括入边、出边和双向边。

Edge Direction

边类型	语法	说明
入边	<-[edge]-	匹配图中指向点的边
出边	-[edge]->	匹配图中从点指出的边
双向边	-[edge]-	匹配图中点的出边和入边

Example

• Basic Mathch

-- Match all node 
MATCH (a)

-- Match all person node
MATCH (a:person)

-- Match node where id = 1
MATCH (a:person where id = 1)

-- One hop match
MATCH (a:person where id = 1)-[e:knows where e.weight > 0.4]->(b:person)

-- Tow hop match
MATCH (a:person)-(b:person) <- (c)

-- Match in-vertex for node a
MATCH (a:person)<-[e:knows]-(b)

• Match With Filter

MATCH (a:person)<-[e:knows]-(b) Where a.id = b.id

• Match Join Match two path pattern and join them with the common label. e.g.

MATCH (a) -> (b), (a) -> (c)

The output is p1 = (a, b) join p2 = (a, c) on p1.a = p2.a.

Regex-Match

不定跳数匹配，类似正则表达式写法指定跳数。

Syntax

PathPatthern: Node Edge '{' minHop ',' [ maxHop] '}' Node

Example

MATCH (a) -[e]->{1,5} (b)

MATCH (a) -[e]->{1,}  (b)

Return

Syntax

RETURN expr {',' expr}* 
[ GROUP BY expr {',' expr}* ] 
[ ORDER BY expr [ASC|DESC] {',' expr [ASC|DESC]} ]
[ LIMIT number ]

Example

MATCH (a:person WHERE a.id = '1')-[e:knows]->(b:person)
RETURN a.name as name, b.id as b_id

MATCH (a:person WHERE a.id = '1')-[e:knows]->(b:person)
RETURN a, b

-- GROUP BY
MATCH (a:person)-[e:knows where e.weight > 0.4]->(b:person)
RETURN a.id, SUM(e.weight) * 10 as amt GROUP BY a.id

-- ORDER BY
MATCH (a:person WHERE a.id = '1')-[e:knows]->(b:person)
RETURN a, b order by a.age DESC, b.age ASC

-- LIMIT
MATCH (a:person WHERE a.id = '1')-[e:knows]->(b:person)
RETURN a, b order by a.age DESC, b.age ASC LIMIT 10

Let

Let语句用于在图匹配过程中修改路径上的点和边的属性值。

Syntax

LET Identifier '.' Identifier = expr

Example

 MATCH (a:person where a.id = 1) -[e:knows]->(b:person)
 LET a.weight = a.age / cast(100.0 as double),
 LET b.weight = b.age / cast(100.0 as double)


MATCH (a:person where a.id = 1) -[e:knows]->(b:person)
LET a.weight = a.age / cast(100.0 as double),
LET a.weight = a.weight * 2,
LET b.weight = 1.0,
LET b.age = 20

SubQuery

Syntax

Scalar Query

AggregateFunction '(' PathPatthern '=>' expr ')'

Exists Query

EXISTS PathPatthern

Example

Scalar Query Example

MATCH (a:person WHERE id = 1)-[e]->(b)
Where COUNT((b) ->(c) => c) >= 1
RETURN a, e, b

MATCH (a:person WHERE id = 1)-[e]->(b)
Let b.out_cnt = COUNT((b) ->(c) => c),
Let b.out_weight = SUM((b) -[e1]-> (c) => e1.weight)
RETURN a, e, b

Exists Query Example

MATCH (a:person WHERE id = 1)-[e]->(b)
Where EXISTS (b) -> (c)
      And SUM((b) -[e1]-> (c) => e1.weight) > 1
RETURN a, e, b

Continue-Match

用户可以将一段复杂的Match拆分成多段Match，多段Match的结果为各个Match路径的join关联。

Syntax

MatchStatement
MatchStatement

Example

MATCH (a:person where a.id = 1) -[e:knows]->(b:person)
LET a.weight = a.age / cast(100.0 as double),
LET a.weight = a.weight * 2
MATCH(b) -[]->(c)
RETURN a.id as a_id, a.weight as a_weight, b.id as b_id, c.id as c_id


MATCH (a) -> (b) where b.id > 0 and a.lang is null
MATCH (a) <- (c) where label(c) = 'person'
Let c.kind = 'k' || cast(c.age / 10 as varchar)
MATCH (c) -> (d) where d != b
Let d.type = if (label(d) = 'person', 1, 0)
RETURN a, b, c, d