MySQL在一个查询中选择多个计数

Question

我有 2 张桌子 - departments 和 employees。我想从departments 中选择所有列，并添加一个列，其中包含子部门数和每个部门的员工数。

简化后的表格如下所示：

departments:

id      name            parent_id
---------------------------------
1       IT              NULL
2       HR              NULL
3       Bussiness       NULL
4       Web dev         1
5       Ecommerce       4
6       Advertisement   3
7       Control         3
8       Programmers     1

employees:

id      name            department_id
---------------------------------
1       Adam            1
2       Ben             8
3       Charles         7
4       David           4
5       Eugen           4
6       Frank           6
7       Gustav          6
8       Heremy          4
9       Igor            5
10      Jacob           3

我想要的结果是：

id      name            parent_id       department_count      employee_count
----------------------------------------------------------------------------
1       IT              NULL            2                     1
2       HR              NULL            0                     0
3       Bussiness       NULL            2                     1
4       Web dev         1               1                     3
5       Ecommerce       4               0                     1
6       Advertisement   3               0                     2
7       Control         3               0                     1
8       Programmers     1               0                     1

这是我的查询，但它不能正常工作。但是，如果我删除了COUNT 中的 1 个，则数字是正确的。我认为NULL 值是问题所在，但我不知道解决方案。

SELECT d.*, COUNT(d2.parent_id) AS department_count, COUNT(e.id) AS employee_count
FROM department AS d
LEFT JOIN department AS d2 ON (d.id = d2.parent_id)
LEFT JOIN employees AS z ON (e.department_id = d.id)
GROUP BY d.id
ORDER BY d.id

欢迎对我的代码进行任何改进。

Answer 1

在你的计数函数中使用d2.id与distinct，这样一个部门的重复孩子将只计算一次

SELECT d.*,
COUNT(distinct d2.id) AS department_count,
COUNT(distinct e.id) AS employee_count
FROM department AS d
LEFT JOIN department AS d2 ON (d.id = d2.parent_id)
LEFT JOIN employees AS e ON (e.department_id = d.id)
GROUP BY d.id
ORDER BY d.id

Demo

Answer 2

您可以尝试相关查询，因为它们提供了简单的逻辑：

SELECT d.*, 
  (SELECT COUNT(*) FROM departments WHERE parent_id = d.id) as department_count,
  (SELECT COUNT(*) FROM employees WHERE department_id = d.id) as employees_count
FROM departments d