Postgresql 查找每个部门的最高工资-是否有更有效的查询？

我有一个有效的查询，但我怀疑我这样做效率低下。是否有一种更优雅的方法来确定每个部门的最高工资以及获得最高工资的员工

我正在做一个cte，以找到每个部门id的最高工资，然后通过匹配工资和部门id将其与员工数据连接起来。我有下面的代码来构建/填充表和最后的查询

CREATE TABLE employee (
emplid SERIAL PRIMARY KEY,
name VARCHAR NOT NULL,
salary FLOAT NOT NULL,
depid INTEGER
);

INSERT INTO employee (name, salary, depid)
VALUES
('Chris',23456.99,1),
('Bob',98756.34,1),
('Malin',34567.22,2),
('Lisa',34967.73,2),
('Deepak',88582.22,3),
('Chester',99487.41,3);

CREATE TABLE department (
depid SERIAL PRIMARY KEY,
deptname VARCHAR NOT NULL
);

INSERT INTO department (deptname)
VALUES
('Engineering'),
('Sales'),
('Marketing');

--top salary by department
WITH cte AS (
SELECT d.depid, deptname, MAX(salary) AS maxsal
FROM employee e
JOIN department d ON d.depid = e.depid
GROUP BY d.depid, deptname
)
SELECT cte.deptname, e.name, cte.maxsal
FROM cte
JOIN employee e ON cte.depid = e.depid
AND e.salary = cte.maxsal
ORDER BY maxsal DESC;

目标结果如下：

“营销”“切斯特”“99487.41” “工程”“Bob”“98756.34”

---

“Sales”“Lisa”“34967.73”

在Postgres中，可以使用

distinct on（）

操作符解决此问题：

SELECT distinct on (d.depid) d.depid, deptname, e.name, e.salary AS maxsal
FROM employee e
  JOIN department d ON d.depid = e.depid
order by d.depid, e.salary desc;

或者您可以使用窗口功能：

select depid, deptname, emp_name, salary
from (
  SELECT d.depid, 
         deptname, 
         e.name as emp_name, 
         e.salary, 
         max(e.salary) over (partition by d.depid) AS maxsal
  FROM employee e
    JOIN department d ON d.depid = e.depid
) t
where salary = maxsal;

---

在线示例：

您应该有一个索引：

create index employee_depid_salary_desc_idx on employee(depid, salary desc);

然后使用以下可以使用索引的查询：

select
  depid,
  deptname,
  (
    select
      emplid
    from employees
    where depid=department.depid
    order by salary desc
    limit 1
  ) as max_salaried_emplid
from department;

---

（用于从EmployId检索数据的联接留给读者作为练习）。

相关：警告：如果一个部门内有两名或两名以上员工的最高工资，则上述查询将检索其中一名员工，但未定义是哪一名员工。