图数据库NebulaGraph学习

1.图空间(Space)操作

1.1创建图空间，指定vid_type为整形

CREATE SPACE play_space (partition_num = 10, replica_factor = 1, vid_type = INT64) COMMENT = "运动员库表空间";

1.2创建图空间，指定vid_type为字符串

CREATE SPACE play_space (partition_num = 10, replica_factor = 1, vid_type = FIXED_STRING(32)) COMMENT = "运动员库表空间";

1.3删除图空间

drop SPACE play_space;

2.标签(Tag)操作

2.1创建点

CREATE tag player (name string NULL ?COMMENT "姓名", age int8 NULL ?COMMENT "年龄") ?COMMENT = "球员";
CREATE tag team (name string NULL ?COMMENT "球队名称") ?COMMENT = "归属球队";

2.2查看库下面的点

SHOW TAGS;

2.2查看某个点的属性

describe TAG player;

2.3基于标签创建点

单条插入示例:

insert vertex player(name,age) values 100:("詹姆斯",38);
insert vertex player(name,age) values 101:("库里",34);
insert vertex player(name,age) values 102:("杜兰特",34);

批量插入示例:

insert vertex team(name) values 200:("洛杉矶湖人队"),201:("金州勇士队"),202:("布鲁克林篮网队");

2.4查看点

指定属性查看示例:

fetch prop on player 100,101,102 YIELD properties(vertex).name,properties(vertex).age;

查询全部属性示例?:

fetch prop on team 200,201,202 YIELD properties(vertex);

2.5删除点

delete vertex 100;

3.边(Edge)操作

3.1创建边

CREATE edge follow (attention_rate int16 NULL ?COMMENT "关注度") ?COMMENT = "球员关注度关联表";
CREATE edge service (start_time int32 NULL ?COMMENT "开始时间", end_time int32 NULL ?COMMENT "结束时间") ?COMMENT = "球员服役时间关联表";

3.2查看库下面的边

SHOW edges;

3.3查看边的属性

describe edge service;

3.4插入边

单条插入示例:

insert edge follow(attention_rate) values 100->101:(20);
insert edge follow(attention_rate) values 101->102:(50);
insert edge follow(attention_rate) values 102->103:(80);

批量插入示例:

insert edge service(start_time,end_time) values 100->200:(2003,2023),101->201:(2009,2023),102->203:(2007,2023);

3.5查看边

单条查看示例:

fetch prop on follow 100->101 YIELD properties(edge);

批量查看示例:

fetch prop on service 100->200,101->201 YIELD properties(edge).start_time,properties(edge).end_time;

4.查看图谱关系图

5.索引创建

5.1创建索引

CREATE {TAG | EDGE} INDEX [IF NOT EXISTS] <index_name>
ON {<tag_name> | <edge_name>} ([<prop_name_list>]) [COMMENT = '<comment>'];

示例

//点
CREATE TAG INDEX IF NOT EXISTS index_player ON player(name(30), age);
//边
CREATE EDGE INDEX IF NOT EXISTS index_follow ON follow(attention_rate);

5.2重建索引

REBUILD {TAG | EDGE} INDEX <index_name>;

示例

//点
REBUILD TAG INDEX index_name;
//边
REBUILD EDGE INDEX index_follow;

6.常用查询

6.1Go语句查询

????????语句从一个或多个点开始，沿着一条或多条边遍历，返回YIELD子句中指定的信息。

GO [[<M> TO] <N> STEPS ] FROM <vertex_list>
OVER <edge_type_list> [{REVERSELY | BIDIRECT}]
[ WHERE <conditions> ]
YIELD [DISTINCT] <return_list>
[{ SAMPLE <sample_list> | <limit_by_list_clause> }]
[| GROUP BY {<col_name> | expression> | <position>} YIELD <col_name>]
[| ORDER BY <expression> [{ASC | DESC}]]
[| LIMIT [<offset>,] <number_rows>];

从vid为100的球员开始找，沿着service边找到终点vid，输出起点vid和终点vid及其信息

GO FROM 100 OVER service YIELD id($^),properties($^),id($$),properties($$);

备注:

$^ :表示边的起点
$$ :表示边的终点
properties() :返回点/边的所有属性

6.2FETCH语句查询

可以获得点或边的属性

6.2.1查看点

查看点

语法:
FETCH PROP ON {<tag_name>[, tag_name ...] | *}
<vid> [, vid ...]
YIELD <return_list> [AS <alias>];

示例:

fetch prop on player 100 YIELD properties(vertex)；

6.2.2查看边

语法:
FETCH PROP ON <edge_type> <src_vid> -> <dst_vid>[@<rank>] [, <src_vid> -> <dst_vid> ...]
YIELD <output>;

示例:

fetch prop on follow 100->101 YIELD properties(edge);

6.3LOOKUP查找

????????语句是基于索引的，和WHERE子句一起使用，查找符合特定条件的数据，要先创建索引

LOOKUP ON {<vertex_tag> | <edge_type>}
[WHERE <expression> [AND <expression> ...]]
YIELD <return_list> [AS <alias>];

<return_list>
    <prop_name> [AS <col_alias>] [, <prop_name> [AS <prop_alias>] ...];

a.找出点中名称为詹姆斯的球员

LOOKUP ON player WHERE player.name == "詹姆斯" YIELD properties(vertex);?

?

b.找出球员中名字经库里开始，并且年龄在34和38两个值之前的球员

LOOKUP ON player WHERE player.name STARTS WITH "库里" AND player.age IN [34,38] YIELD properties(vertex).name, properties(vertex).age;

?

c.根据条件查看边

LOOKUP ON follow WHERE follow.attention_rate > 20 YIELD properties(edge);?

d.统计边或者点的数量?

LOOKUP ON player YIELD id(vertex)| YIELD COUNT(*) AS Player_Number;
LOOKUP ON follow YIELD edge AS e| YIELD COUNT(*) AS Follow_Number;

?

6.4MATCH查找

????????可以灵活的描述各种图模式，但是它依赖索引去匹配 Nebula Graph 中的数据模型，性能也还需要调优，要先创建索引

MATCH <pattern> [<clause_1>] ?RETURN <output> ?[<clause_2>];

查找球员名称为的点并返回

?MATCH (v:player{name:"杜兰特"}) RETURN v