跳转至

GROUP BYGraph

GROUP BY子句可以用于聚合数据。

openCypher兼容性Graph

本文操作仅适用于原生nGQL。

用户也可以使用openCypher方式的Graph函数聚合数据。

nebula>  MATCH (v:player)<-[:follow]-(:player) RETURN v.name AS Name, count(*) as cnt ORDER BY cnt DESC;
+----------------------+-----+
| Name                 | cnt |
+----------------------+-----+
| "Tim Duncan"         | 10  |
| "LeBron James"       | 6   |
| "Tony Parker"        | 5   |
| "Chris Paul"         | 4   |
| "Manu Ginobili"      | 4   |
+----------------------+-----+
...

语法Graph

GROUP BY子句可以聚合相同值的行，然后进行计数、排序和计算等操作。

GROUP BY子句可以在管道符（|）之后和YIELD子句之前使用。

| GROUP BY <var> YIELD <var>, <aggregation_function(var)>
aggregation_function()函数支持avg()、sum()、max()、min()、count()、collect()、std()。

示例Graph

# 查找所有连接到player100的点，并根据他们的姓名进行分组，返回姓名的出现次数。
nebula> GO FROM "player100" OVER follow BIDIRECT \
        YIELD properties($$).name as Name \
        | GROUP BY $-.Name \
        YIELD $-.Name as Player, count(*) AS Name_Count;
+---------------------+------------+
| Player              | Name_Count |
+---------------------+------------+
| "Shaquille O'Neal"  | 1          |
| "Tiago Splitter"    | 1          |
| "Manu Ginobili"     | 2          |
| "Boris Diaw"        | 1          |
| "LaMarcus Aldridge" | 1          |
| "Tony Parker"       | 2          |
| "Marco Belinelli"   | 1          |
| "Dejounte Murray"   | 1          |
| "Danny Green"       | 1          |
| "Aron Baynes"       | 1          |
+---------------------+------------+

用函数进行分组和计算Graph

# 查找所有连接到player100的点，并根据起始点进行分组，返回degree的总和。
nebula> GO FROM "player100" OVER follow \
        YIELD src(edge) AS player, properties(edge).degree AS degree \
        | GROUP BY $-.player \
        YIELD sum($-.degree);
+----------------+
| sum($-.degree) |
+----------------+
| 190            |
+----------------+

sum()函数详情请参见Graph。

最后更新: November 1, 2021
Back to top