步骤4：使用常用nGQL（CRUD命令）

本文介绍Nebula Graph查询语言的基础语法，包括用于Schema创建和常用增删改查操作的语句。

如需了解更多语句的用法，参见nGQL指南。

图空间和Schema

一个Nebula Graph实例由一个或多个图空间组成。每个图空间都是物理隔离的，用户可以在同一个实例中使用不同的图空间存储不同的数据集。

Nebula Graph and graph spaces

为了在图空间中插入数据，需要为图数据库定义一个Schema。Nebula Graph的Schema是由如下几部分组成。

组成部分说明
点（Vertex）表示现实世界中的实体。一个点可以有一个或多个标签。
标签（Tag）点的类型，定义了一组描述点类型的属性。
边（Edge）表示两个点之间有方向的关系。
边类型（Edge type）边的类型，定义了一组描述边的类型的属性。

更多信息，请参见数据结构。

本文将使用下图的数据集演示基础操作的语法。

The demo dataset

检查Nebula Graph集群的机器状态

  1. nebula> SHOW HOSTS;
  2. +-------------+-----------+-----------+--------------+----------------------+------------------------+
  3. | Host        | Port      | Status    | Leader count | Leader distribution  | Partition distribution |
  4. +-------------+-----------+-----------+--------------+----------------------+------------------------+
  5. | "storaged0" | 9779      | "ONLINE"  | 0            | "No valid partition" | "No valid partition"   |
  6. +-------------+-----------+-----------+--------------+----------------------+------------------------+
  7. | "storaged1" | 9779      | "ONLINE"  | 0            | "No valid partition" | "No valid partition"   |
  8. +-------------+-----------+-----------+--------------+----------------------+------------------------+
  9. | "storaged2" | 9779      | "ONLINE"  | 0            | "No valid partition" | "No valid partition"   |
  10. +-------------+-----------+-----------+--------------+----------------------+------------------------+
  11. | "Total"     | __EMPTY__ | __EMPTY__ | 0            | __EMPTY__            | __EMPTY__              |
  12. +-------------+-----------+-----------+--------------+----------------------+------------------------+
  13. Got 4 rows (time spent 1061/2251 us)

在返回结果中，查看Status列，可以看到所有Storage服务都在线。

异步实现创建和修改

  • CREATE SPACE
  • CREATE TAG
  • CREATE EDGE
  • ALTER TAG
  • ALTER EDGE
  • CREATE TAG INDEX
  • CREATE EDGE INDEX

为确保数据同步，后续操作能顺利进行，可采取以下方法之一：

  • 执行SHOW或DESCRIBE命令检查相应对象的状态，确保创建或修改已完成。如果没有完成，请等待几秒重试。

  • 等待2个心跳周期（20秒）。

创建和选择图空间

nGQL语法

  • 创建图空间

    1. CREATE SPACE [IF NOT EXISTS] <graph_space_name> (
    2. [partition_num = <partition_number>,] 
    3. [replica_factor = <replica_number>,]
    4. vid_type = {FIXED_STRING(<N>) | INT64}
    5. )
    6. [COMMENT = '<comment>'];

    参数详情请参见CREATE SPACE。

  • 列出创建成功的图空间

    1. nebula> SHOW SPACES;

  • 选择数据库

    1. USE <graph_space_name>;

示例

  1. 执行如下语句创建名为basketballplayer的图空间。

    1. nebula> CREATE SPACE basketballplayer(partition_num=15, replica_factor=1, vid_type=fixed_string(30));
    2. Execution succeeded (time spent 2817/3280 us)

  2. 执行命令SHOW HOSTS检查分片的分布情况，确保平衡分布。

    1. nebula> SHOW HOSTS;
    2. +-------------+-----------+-----------+--------------+----------------------------------+------------------------+
    3. | Host        | Port      | Status    | Leader count | Leader distribution              | Partition distribution |
    4. +-------------+-----------+-----------+--------------+----------------------------------+------------------------+
    5. | "storaged0" | 9779      | "ONLINE"  | 5            | "basketballplayer:5"             | "basketballplayer:5"   |
    6. +-------------+-----------+-----------+--------------+----------------------------------+------------------------+
    7. | "storaged1" | 9779      | "ONLINE"  | 5            | "basketballplayer:5"             | "basketballplayer:5"   |
    8. +-------------+-----------+-----------+--------------+----------------------------------+------------------------+
    9. | "storaged2" | 9779      | "ONLINE"  | 5            | "basketballplayer:5"             | "basketballplayer:5"   |
    10. +-------------+-----------+-----------+--------------+----------------------------------+------------------------+
    11. | "Total"     |           |           | 15           | "basketballplayer:15"            | "basketballplayer:15"  |
    12. +-------------+-----------+-----------+--------------+----------------------------------+------------------------+
    13. Got 4 rows (time spent 1633/2867 us)

    如果Leader distribution分布不均匀，请执行命令BALANCE LEADER重新分配。更多信息，请参见Storage负载均衡。

  3. 选择图空间basketballplayer。

    1. nebula[(none)]> USE basketballplayer;
    2. Execution succeeded (time spent 1229/2318 us)

    用户可以执行命令SHOW SPACES查看创建的图空间。

    1. nebula> SHOW SPACES;
    2. +--------------------+
    3. | Name               |
    4. +--------------------+
    5. | "basketballplayer" |
    6. +--------------------+
    7. Got 1 rows (time spent 977/2000 us)

创建Tag和Edge type

nGQL语法

  1. CREATE {TAG | EDGE} {<tag_name> | <edge_type>}(<property_name> <data_type>
  2. [, <property_name> <data_type> ...])
  3. [COMMENT = '<comment>'];

参数详情请参见CREATE TAG和CREATE EDGE。

示例

创建Tag:player和team，以及Edge type:follow和serve。说明如下表。

名称类型属性
playerTagname (string), age (int)
teamTagname (string)
followEdge typedegree (int)
serveEdge typestart_year (int), end_year (int)
  1. nebula> CREATE TAG player(name string, age int);
  2. Execution succeeded (time spent 20708/22071 us)
  4. nebula> CREATE TAG team(name string);
  5. Execution succeeded (time spent 5643/6810 us)
  7. nebula> CREATE EDGE follow(degree int);
  8. Execution succeeded (time spent 12665/13934 us)
  10. nebula> CREATE EDGE serve(start_year int, end_year int);
  11. Execution succeeded (time spent 5858/6870 us)

插入点和边

用户可以使用INSERT语句，基于现有的Tag插入点，或者基于现有的Edge type插入边。

nGQL语法

  • 插入点

    1. INSERT VERTEX [IF NOT EXISTS] <tag_name> (<property_name>[, <property_name>...])
    2. [, <tag_name> (<property_name>[, <property_name>...]), ...]
    3. {VALUES | VALUE} <vid>: (<property_value>[, <property_value>...])
    4. [, <vid>: (<property_value>[, <property_value>...];

    VID是Vertex ID的缩写，VID在一个图空间中是唯一的。参数详情请参见INSERT VERTEX。

  • 插入边

    1. INSERT EDGE [IF NOT EXISTS] <edge_type> (<property_name>[, <property_name>...])
    2. {VALUES | VALUE} <src_vid> -> <dst_vid>[@<rank>] : (<property_value>[, <property_value>...])
    3. [, <src_vid> -> <dst_vid>[@<rank>] : (<property_name>[, <property_name>...]), ...];

    参数详情请参见INSERT EDGE。

示例

  • 插入代表球员和球队的点。

    1. nebula> INSERT VERTEX player(name, age) VALUES "player100":("Tim Duncan", 42);
    2. Execution succeeded (time spent 28196/30896 us)
    4. nebula> INSERT VERTEX player(name, age) VALUES "player101":("Tony Parker", 36);
    5. Execution succeeded (time spent 2708/3834 us)
    7. nebula> INSERT VERTEX player(name, age) VALUES "player102":("LaMarcus Aldridge", 33);
    8. Execution succeeded (time spent 1945/3294 us)
    10. nebula> INSERT VERTEX team(name) VALUES "team200":("Warriors"), "team201":("Nuggets");
    11. Execution succeeded (time spent 2269/3310 us)

  • 插入代表球员和球队之间关系的边。

    1. nebula> INSERT EDGE follow(degree) VALUES "player100" -> "player101":(95);
    2. Execution succeeded (time spent 3362/4542 us)
    4. nebula> INSERT EDGE follow(degree) VALUES "player100" -> "player102":(90);
    5. Execution succeeded (time spent 2974/4274 us)
    7. nebula> INSERT EDGE follow(degree) VALUES "player102" -> "player101":(75);
    8. Execution succeeded (time spent 1891/3096 us)
    10. nebula> INSERT EDGE serve(start_year, end_year) VALUES "player100" -> "team200":(1997, 2016), "player101" -> "team201":(1999,  2018);
    11. Execution succeeded (time spent 6064/7104 us)

查询数据

  • GO语句可以根据指定的条件遍历数据库。GO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。

  • FETCH语句可以获得点或边的属性。

  • LOOKUP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。

  • MATCH语句是查询图数据最常用的，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。

nGQL语法

  • GO

    1. GO [[<M> TO] <N> STEPS ] FROM <vertex_list>
    2. OVER <edge_type_list> [REVERSELY] [BIDIRECT]
    3. [WHERE <expression> [AND | OR expression ...])]
    4. YIELD [DISTINCT] <return_list>;

  • FETCH

    • 查询Tag属性

      1. FETCH PROP ON {<tag_name> | <tag_name_list> | *} <vid_list>
      2. [YIELD [DISTINCT] <return_list>];

    • 查询边属性

      1. FETCH PROP ON <edge_type> <src_vid> -> <dst_vid>[@<rank>]
      2. [, <src_vid> -> <dst_vid> ...]
      3. [YIELD [DISTINCT] <return_list>];

  • LOOKUP

    1. LOOKUP ON {<tag_name> | <edge_type>} 
    2. WHERE <expression> [AND expression ...])]
    3. [YIELD <return_list>];

  • MATCH

    1. MATCH <pattern> [<WHERE clause>] RETURN <output>;

GO语句示例

  • 从VID为player100的球员开始，沿着边follow找到连接的球员。

    1. nebula> GO FROM "player100" OVER follow;
    2. +-------------+
    3. | follow._dst |
    4. +-------------+
    5. | "player101" |
    6. +-------------+
    7. | "player102" |
    8. +-------------+
    9. Got 2 rows (time spent 12097/14220 us)

  • 从VID为player100的球员开始，沿着边follow查找年龄大于或等于35岁的球员，并返回他们的姓名和年龄，同时重命名对应的列。

    1. nebula> GO FROM "player100" OVER follow WHERE $$.player.age >= 35 \
    2.         YIELD $$.player.name AS Teammate, $$.player.age AS Age;
    3. +---------------+-----+
    4. | Teammate      | Age |
    5. +---------------+-----+
    6. | "Tony Parker" | 36  |
    7. +---------------+-----+
    8. Got 1 rows (time spent 8206/9335 us)
    子句/符号说明
    YIELD指定该查询需要返回的值或结果。
    $$表示边的终点。
    \表示换行继续输入。

  • 从VID为player100的球员开始，沿着边follow查找连接的球员，然后检索这些球员的球队。为了合并这两个查询请求，可以使用管道符或临时变量。

    • 使用管道符

      1. nebula> GO FROM "player100" OVER follow YIELD follow._dst AS id | \
      2.         GO FROM $-.id OVER serve YIELD $$.team.name AS Team, \
      3.         $^.player.name AS Player;
      4. +-----------+---------------+
      5. | Team      | Player        |
      6. +-----------+---------------+
      7. | "Nuggets" | "Tony Parker" |
      8. +-----------+---------------+
      9. Got 1 rows (time spent 5055/8203 us)
      子句/符号说明
      $^表示边的起点。
      |组合多个查询的管道符，将前一个查询的结果集用于后一个查询。
      $-表示管道符前面的查询输出的结果集。

    • 使用临时变量

      1. nebula> $var = GO FROM "player100" OVER follow YIELD follow._dst AS id; \
      2.         GO FROM $var.id OVER serve YIELD $$.team.name AS Team, \
      3.         $^.player.name AS Player;
      4. +---------+-------------+
      5. | Team    | Player      |
      6. +---------+-------------+
      7. | Nuggets | Tony Parker |
      8. +---------+-------------+
      9. Got 1 rows (time spent 3103/3711 us)

FETCH语句示例

查询VID为player100的球员的属性。

  1. nebula> FETCH PROP ON player "player100";
  2. +----------------------------------------------------+
  3. | vertices_                                          |
  4. +----------------------------------------------------+
  5. | ("player100" :player{age: 42, name: "Tim Duncan"}) |
  6. +----------------------------------------------------+
  7. Got 1 rows (time spent 2006/2406 us)

修改点和边

用户可以使用UPDATE语句或UPSERT语句修改现有数据。

UPSERT是UPDATE和INSERT的结合体。当使用UPSERT更新一个点或边，如果它不存在，数据库会自动插入一个新的点或边。

nGQL语法

  • UPDATE点

    1. UPDATE VERTEX <vid> SET <properties to be updated>
    2. [WHEN <condition>] [YIELD <columns>];

  • UPDATE边

    1. UPDATE EDGE <source vid> -> <destination vid> [@rank] OF <edge_type>
    2. SET <properties to be updated> [WHEN <condition>] [YIELD <columns to be output>];

  • UPSERT点或边

    1. UPSERT {VERTEX <vid> | EDGE <edge_type>} SET <update_columns>
    2. [WHEN <condition>] [YIELD <columns>];

示例

  • 用UPDATE修改VID为player100的球员的name属性，然后用FETCH语句检查结果。

    1. nebula> UPDATE VERTEX "player100" SET player.name = "Tim";
    2. Execution succeeded (time spent 3483/3914 us)
    4. nebula> FETCH PROP ON player "player100";
    5. +---------------------------------------------+
    6. | vertices_                                   |
    7. +---------------------------------------------+
    8. | ("player100" :player{age: 42, name: "Tim"}) |
    9. +---------------------------------------------+
    10. Got 1 rows (time spent 2463/3042 us)

  • 用UPDATE修改某条边的degree属性，然后用FETCH检查结果。

    1. nebula> UPDATE EDGE "player100" -> "player101" OF follow SET degree = 96;
    2. Execution succeeded (time spent 3932/4432 us)
    4. nebula> FETCH PROP ON follow "player100" -> "player101";
    5. +----------------------------------------------------+
    6. | edges_                                             |
    7. +----------------------------------------------------+
    8. | [:follow "player100"->"player101" @0 {degree: 96}] |
    9. +----------------------------------------------------+
    10. Got 1 rows (time spent 2205/2800 us)

  • 用UPSERT插入一个VID为player111的点。

    1. nebula> INSERT VERTEX player(name, age) VALUES "player111":("Ben Simmons", 22);
    2. Execution succeeded (time spent 2115/2900 us)
    4. Wed, 21 Oct 2020 11:11:50 UTC
    6. nebula> UPSERT VERTEX "player111" SET player.name = "Dwight Howard", player.age = $^.player.age + 11 \
    7.         WHEN $^.player.name == "Ben Simmons" AND $^.player.age > 20 \
    8.         YIELD $^.player.name AS Name, $^.player.age AS Age;
    9. +---------------+-----+
    10. | Name          | Age |
    11. +---------------+-----+
    12. | Dwight Howard | 33  |
    13. +---------------+-----+
    14. Got 1 rows (time spent 1815/2329 us)

删除点和边

nGQL语法

  • 删除点

    1. DELETE VERTEX <vid1>[, <vid2>...]

  • 删除边

    1. DELETE EDGE <edge_type> <src_vid> -> <dst_vid>[@<rank>]
    2. [, <src_vid> -> <dst_vid>...]

示例

  • 删除点

    1. nebula> DELETE VERTEX "team1", "team2";
    2. Execution succeeded (time spent 4337/4782 us)

  • 删除边

    1. nebula> DELETE EDGE follow "team1" -> "team2";
    2. Execution succeeded (time spent 3700/4101 us)

索引

用户可以通过CREATE INDEX语句为Tag和Edge type增加索引。

nGQL语法

  • 创建索引

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

  • 重建索引

    1. REBUILD {TAG | EDGE} INDEX <index_name>;

示例

为Tagplayer的属性name创建索引，并且重建索引。

  1. nebula> CREATE TAG INDEX player_index_0 on player(name(20));
  2. nebula> REBUILD TAG INDEX player_index_0;

基于索引的LOOKUP和MATCH示例

确保LOOKUP或MATCH有一个索引可用。如果没有，请先创建索引。

找到Tag为player的点的信息，它的name属性值为Tony Parker。

  1. // 为name属性创建索引player_name_0。
  2. nebula> CREATE TAG INDEX player_name_0 on player(name(10));
  3. Execution succeeded (time spent 3465/4150 us)
  5. // 重建索引确保能对已存在数据生效。
  6. nebula> REBUILD TAG INDEX player_name_0
  7. +------------+
  8. | New Job Id |
  9. +------------+
  10. | 31         |
  11. +------------+
  12. Got 1 rows (time spent 2379/3033 us)
  14. // 使用LOOKUP语句检索点的属性。
  15. nebula> LOOKUP ON player WHERE player.name == "Tony Parker" \
  16.         YIELD player.name, player.age;
  17. +-------------+---------------+------------+
  18. | VertexID    | player.name   | player.age |
  19. +-------------+---------------+------------+
  20. | "player101" | "Tony Parker" | 36         |
  21. +-------------+---------------+------------+
  23. // 使用MATCH语句检索点的属性。
  24. nebula> MATCH (v:player{name:"Tony Parker"}) RETURN v;
  25. +-----------------------------------------------------+
  26. | v                                                   |
  27. +-----------------------------------------------------+
  28. | ("player101" :player{age: 36, name: "Tony Parker"}) |
  29. +-----------------------------------------------------+
  30. Got 1 rows (time spent 5132/6246 us)