如何在mysql中联合后获取最小值？

Question

假设我有 2 个表 A 和 B。

这 2 个表共有 3 列，分别是 Name、Id 和 Price。

这是我用于 1 个表的查询：

SELECT Name, Id, Price FROM A WHERE Id = "123" and Price = (SELECT MIN(Price) FROM A);

我刚刚意识到，当最低价格由另一个 Id 持有时，此查询不起作用。

所以我环顾四周，我认为我应该使用 GROUP BY ？

我已将其更改为：

SELECT Name, MIN(Price) FROM A WHERE Id = "123" GROUP BY Name;

但这不是预期的结果。

假设在表 A 中我有：

Name	Id	Price
Au	123	12
Be	123	16
St	122	9
Ge	123	10

对于表 B，我有：

Name	Id	Price
La	123	14.5
La	123	12
St	123	13
Is	123	12
Is	123	10
La	123	10
Is	123	10

而预期的结果是：

Name	Price
Ge	10
Is	10
La	10

预期结果是 1 行长，因为在数据集中只有一行与条件匹配，但如果我有另一行价格为 10 且 ID 为 123，它也应该存在。因此，如果有更多符合条件的行，我希望它们出现在结果中。

问题是，当我使用 UNION 执行以下查询时，我不知道如何获得特定 ID 的最低价格：

SELECT Name, Id, Price FROM A UNION SELECT Name, Id, Price FROM B;

那么我可以在查询中添加什么以获得预期的结果，然后如果我使用联合来获得 2 个表中特定 ID 的最低价格，它将如何工作？

Answer 1

在 MySQL 8+ 上，我们可以在这里使用 RANK 和联合查询：

WITH cte AS (
    SELECT Name, Id, Price FROM A WHERE Id = '123'
    UNION ALL
    SELECT Name, Id, Price FROM B WHERE Id = '123'
),
cte2 AS (
    SELECT *, RANK() OVER (ORDER BY price) rnk
    FROM cte
)

SELECT Name, Id, Price
FROM cte2
WHERE rnk = 1;

这是一个适用于早期版本的 MySQL 的查询：

SELECT Name, Id, Price
FROM
(
    SELECT Name, Id, Price FROM A WHERE Id = '123'
    UNION ALL
    SELECT Name, Id, Price FROM B WHERE Id = '123'
) t
WHERE Price = (
    SELECT Price FROM A WHERE Id = '123'
    UNION ALL
    SELECT Price FROM B WHERE Id = '123'
    ORDER BY Price
    LIMIT 1
);

Answer 2

这对你来说应该没问题：

select * 
from (select Name, id, Price FROM A
      union
      select Name, id, Price FROM B) Tab
where (Tab.id, Tab.price) = (select Tab2.id, min(Tab2.price)
                             from (select Name, id, Price FROM A
                                   union
                                   select Name, id, Price FROM B) Tab2
                             where Tab2.id = '123')

您只有一个地方可以放置您要查找的 ID。

在这里你可以看到演示： DEMO

/*This returns everything from your two tables*/
select * 
from (SELECT Name, id, Price FROM A
      union
      select Name, id, Price FROM B) Tab

/*this returns the minimal price for your requested ID, here you requested id =123*/
select Tab2.id, min(Tab2.price)
from (SELECT Name, id, Price FROM A
      union
      select Name, id, Price FROM B) Tab2
where Tab2.id = '123'

--with this where clause: 
where (Tab.id, Tab.price)
/*you are telling the query :
  give me every row from all the data(first query)that has 
  this combination of ID + PRICE: 
  123 + 10 (you have found this with the second query)

  So, you do not care what name it is, it only has to have :
  ID = 123 and the lowest price which is 10.
  ID 123 was requested from you and lowest price for that ID is 10,  
  which you have founded with the second query.*/

Answer 3

我在 db-fiddle.com 上对此进行了测试，它返回所有价格最低的行：

SELECT Id, Name, Price
FROM (SELECT * FROM A UNION SELECT * FROM B) TMP
WHERE (Price, Id) = (
    SELECT MIN(Price), Id
    FROM (SELECT * FROM A UNION SELECT * FROM B) TMP2
    WHERE Id = "123"
);

这是我测试查询所针对的表的脚本：

create table A(
   _id INT NOT NULL AUTO_INCREMENT,
   Id  VARCHAR(100) NOT NULL,
   Name VARCHAR(100) NOT NULL,
   Price INT NOT NULL,
   PRIMARY KEY ( _id )
);

create table B(
   _id INT NOT NULL AUTO_INCREMENT,
   Id  VARCHAR(100) NOT NULL,
   Name VARCHAR(100) NOT NULL,
   Price INT NOT NULL,
   PRIMARY KEY ( _id )
);

INSERT INTO A(Id, Name, Price)
VALUES
('123', 'Name123a1', 21),
('123', 'Name123a2', 41),
('124', 'Name124a', 40);

INSERT INTO B(Id, Name, Price)
VALUES
('123', 'Name123b1', 22),
('123', 'Name123b2', 21),
('124', 'Name124b', 20);

解决方案需要一些时间才能弄清楚，因为我生疏了。感谢 VBoka 帮助我解决了错误。

Answer 4

我会在您最后一次查询后执行此操作，即输出 3 个价格值并且您需要最小值的那个。

• 创建一个 cte
• 使用ROW_NUMBER 创建排名
• 过滤具有该排名列的数据

with data as (
    
select 
    *, ROW_NUMBER () OVER(ORDER BY price ) as rank_ from table
    )

select * from data where rank_ = 1