ISO GQL

1.GQL简介

Graph Query Language(GQL, 图查询语言)是一种国际标准语言，用于属性图查询，该语言建立在SQL的基础上，并整合了现有的openCypher、PGQL、GSQL和G-CORE语言的成熟思想。目前该标准仍然处于草稿阶段。

TuGraph基于ISO GQL (ISO/IEC 39075) Antlr4 语法文件实现了GQL，并做了一些扩展与改造。目前并未完全支持所有的GQL语法，我们会在未来逐步完善。

2.Clauses

类别	子句
Reading clauses	MATCH
	OPTIONAL MATCH
Projecting clauses	RETURN
	NEXT
Reading sub-clauses	WHERE
	ORDER BY
	SKIP
	LIMIT

2.1.MATCH

MATCH子句式是GQL最基础的子句，几乎所有查询都是通过 MATCH展开。

MATCH子句用于指定在图中搜索的匹配模式，用来匹配满足一定条件的点或者路径。

点查询

查询所有点

MATCH (n)
RETURN n

查询特定标签的点

MATCH (n:Person)
RETURN n

通过属性匹配点

MATCH (n:Person{name:'Michael Redgrave'})
RETURN n.birthyear

返回结果

[{"n.birthyear":1908}]

通过过滤条件匹配点

MATCH (n:Person WHERE n.birthyear > 1910)
RETURN n.name LIMIT 2

返回结果

[{"n.name":"Christopher Nolan"},{"n.name":"Corin Redgrave"}]

边查询

出边匹配

MATCH (n:Person WHERE n.birthyear = 1970)-[e]->(m)
RETURN n.name, label(e), m.name

返回结果

[{"label(e)":"BORN_IN","m.name":"London","n.name":"Christopher Nolan"},{"label(e)":"DIRECTED","m.name":null,"n.name":"Christopher Nolan"}]

入边匹配

MATCH (n:Person WHERE n.birthyear = 1939)<-[e]-(m)
RETURN n.name, label(e), m.name

返回结果

[{"label(e)":"HAS_CHILD","m.name":"Rachel Kempson","n.name":"Corin Redgrave"},{"label(e)":"HAS_CHILD","m.name":"Michael Redgrave","n.name":"Corin Redgrave"}]

带过滤条件的边匹配

MATCH (n:Person)-[e:BORN_IN WHERE e.weight > 20]->(m)
RETURN n.name, e.weight, m.name

返回结果

[{"e.weight":20.549999237060547,"m.name":"New York","n.name":"John Williams"},{"e.weight":20.6200008392334,"m.name":"New York","n.name":"Lindsay Lohan"}]

路径匹配

不定跳查询

MATCH (n:Person)-[e]->{2,3}(m:Person)
RETURN m.name LIMIT 2

返回结果

[{"m.name":"Liam Neeson"},{"m.name":"Natasha Richardson"}]

2.2.OPTIONAL MATCH

OPTIONAL MATCH匹配图模式，如果未命中，则返回null。

查询命中

OPTIONAL MATCH (n:Person{name:'Michael Redgrave'})
RETURN n.birthyear

返回结果

[{"n.birthyear":1908}]

查询未命中

OPTIONAL MATCH (n:Person{name:'Redgrave Michael'})
RETURN n.birthyear

返回结果

[{"n.birthyear":null}]

2.3.RETURN

RETURN子句指定返回结果，包括返回点、边、路径、属性等。

返回点

MATCH (n)
RETURN n LIMIT 2

返回结果

[{"n":{"identity":0,"label":"Person","properties":{"birthyear":1910,"name":"Rachel Kempson"}}},{"n":{"identity":1,"label":"Person","properties":{"birthyear":1908,"name":"Michael Redgrave"}}}]

返回边

MATCH (n)-[e]->(m)
RETURN e LIMIT 2

返回结果

[{"e":{"dst":2,"forward":false,"identity":0,"label":"HAS_CHILD","label_id":0,"src":0,"temporal_id":0}},{"e":{"dst":3,"forward":false,"identity":0,"label":"HAS_CHILD","label_id":0,"src":0,"temporal_id":0}}]

返回属性

MATCH (n:Person)
RETURN n.name LIMIT 2

返回结果

[{"n.name":"Christopher Nolan"},{"n.name":"Corin Redgrave"}]

不常见字符串作为变量名

MATCH (`/uncommon variable`:Person)
RETURN `/uncommon variable`.name LIMIT 3

返回结果

[{"`/uncommon variable`.name":"Christopher Nolan"},{"`/uncommon variable`.name":"Corin Redgrave"},{"`/uncommon variable`.name":"Dennis Quaid"}]

列别名

MATCH (n:Person)
RETURN n.name AS nname LIMIT 2

返回结果

[{"nname":"Christopher Nolan"},{"nname":"Corin Redgrave"}]

可选属性

MATCH (n:Person)
RETURN n.age LIMIT 2

返回结果

[{"n.age":null},{"n.age":null}]

其它表达式

MATCH (n:Person)
RETURN n.birthyear > 1970, "I'm a literal", 1 + 2, abs(-2)
LIMIT 2

返回结果

[{"\"I'm a literal\"":"I'm a literal","1 + 2":3,"abs(-2)":2,"n.birthyear > 1970":false},{"\"I'm a literal\"":"I'm a literal","1 + 2":3,"abs(-2)":2,"n.birthyear > 1970":false}]

结果唯一性

MATCH (n)
RETURN DISTINCT label(n) AS label

返回结果

[{"label":"Person"},{"label":"City"},{"label":"Film"}]

2.4.NEXT

NEXT子句用于连接多个子句。

连接MATCH

MATCH (n:Person) WHERE n.birthyear = 1970
RETURN n
NEXT
MATCH (m:Person) WHERE m.birthyear < 1968
RETURN n.name, n.birthyear, m.name LIMIT 2

返回结果

[{"m.name":"Rachel Kempson","n.birthyear":1970,"n.name":"Christopher Nolan"},{"m.name":"Michael Redgrave","n.birthyear":1970,"n.name":"Christopher Nolan"}]

2.5.WHERE

WHERE子句用于过滤记录。

过滤点

MATCH (n:Person WHERE n.birthyear > 1965)
RETURN n.name

返回结果

[{"n.name":"Christopher Nolan"},{"n.name":"Lindsay Lohan"}]

过滤边

MATCH (n:Person WHERE n.birthyear > 1965)-[e:ACTED_IN]->(m:Film)
WHERE e.charactername = 'Halle/Annie'
RETURN m.title

返回结果

[{"m.title":"The Parent Trap"}]

布尔表达式

AND, OR, XOR和 NOT布尔表达式可以用在 WHERE中用来过滤数据。

MATCH (n:Person)
WHERE
	n.birthyear > 1930 AND (n.birthyear < 1950 OR n.name = 'Corin Redgrave')
RETURN n LIMIT 2

返回结果

[{"n":{"identity":3,"label":"Person","properties":{"birthyear":1939,"name":"Corin Redgrave"}}},{"n":{"identity":11,"label":"Person","properties":{"birthyear":1932,"name":"John Williams"}}}]

2.6.ORDER BY

ORDER BY是RETURN的子句，对输出的结果进行排序。

对结果排序

MATCH (n:Person WHERE n.birthyear < 1970)
RETURN n.birthyear AS q
ORDER BY q ASC
LIMIT 5

返回结果

[{"q":1873},{"q":1908},{"q":1910},{"q":1930},{"q":1932}]

2.7.SKIP

SKIP指定结果偏移行数。

未使用SKIP

MATCH (n:Person)
RETURN n.name LIMIT 3

返回结果

[{"n.name":"Christopher Nolan"},{"n.name":"Corin Redgrave"},{"n.name":"Dennis Quaid"}]

使用SKIP

MATCH (n:Person)
RETURN n.name SKIP 1 LIMIT 2

返回结果

[{"n.name":"Corin Redgrave"},{"n.name":"Dennis Quaid"}]

2.8.LIMIT

LIMIT限制结果行数。

使用LIMIT

MATCH (n:Person)
RETURN n.name LIMIT 2;

返回结果

[{"n.name":"Christopher Nolan"},{"n.name":"Corin Redgrave"}]