如何在sql中找到具有多个最大值的组的最大值？

Question

表格必须按部门分组，并且必须返回部门的最大数量

表 A：

Id	Name	Department
1	John Abraham	HR
2	Michael Clarke	HR
3	Roy Thomas	Manager
4	Tom Jose	HR
4	Jerry Pinto	Manager

表 B：

M_Id	Amount
1	5000
2	5000
3	2500
4	1000
4	1500

预期答案

Id	Name	Department	Amount
1	John Abraham	HR	5000
2	Michael Clarke	HR	5000
3	Roy Thomas	Manager	2500

Answer 1

你可以试试这样的：

select main.*

FROM

-- get all information from t1 and amount from t2
(
  select t1.*, t2.amount from t1 inner join t2 on t1.id = m_id
) main

INNER JOIN

-- get max amount by department
(
  select department, max(amount) max_amount from t1 inner join t2 on t1.id = m_id
  group by department
) summary

-- match with main by department and the max amount
on main.department = summary.department
and main.amount = summary.max_amount;

结果

+------+----------------+------------+--------+
| id   | name           | department | amount |
+------+----------------+------------+--------+
|    1 | John Abraham   | HR         |   5000 |
|    2 | Michael Clarke | HR         |   5000 |
|    3 | Roy Thomas     | Manager    |   2500 |
+------+----------------+------------+--------+

示例在这里：https://dbfiddle.uk/?rdbms=mysql_5.7&fiddle=a4cdd94b415df204b0fd967263ba9dc8

解释

由于您想按部门获取最大金额，我们为此创建了一个子查询。我们给它一个别名summary。这给了我们这个：

select department, max(amount)
from t1 inner join t2 on t1.id = m_id group by department;
+------------+-------------+
| department | max(amount) |
+------------+-------------+
| HR         |        5000 |
| Manager    |        2500 |
+------------+-------------+

我们通过合并 t1 和 t2 表来合并您想要报告的数据，并为其命名为 main。这给了我们这个：

select t1.*, t2.amount from t1 inner join t2 on t1.id = m_id;
+------+----------------+------------+--------+
| id   | name           | department | amount |
+------+----------------+------------+--------+
|    1 | John Abraham   | HR         |   5000 | <-- want this
|    2 | Michael Clarke | HR         |   5000 | <-- want this
|    3 | Roy Thomas     | Manager    |   2500 | <-- want this
|    4 | Jerry Pinto    | Manager    |   1000 |
|    4 | Tom Jose       | HR         |   1000 |
+------+----------------+------------+--------+

然后，我们确保根据部门和两个子查询中的金额连接两个子查询。