SQL语句编写求助,两张表联合查询

有两张表如图

想根据B.Name in ('张三','李四') 来获取 8条数据
如图
A表复合主键 ID+Vaule
B表主键Name
A.ID=B.ID
简单说来就是 取出B表同一个ID下 小于Value 的最大数据 和大于Value的最小数据
当B.Value处于 0或者max的时候只取出 一条(满足条件)即可

不限制 sql或者存储过程...
只需得到结果

这个问题我已经思考很久了, 并非一点sql 都不会 万望回答的朋友三思

---------------------------------------------------------分割线 以下内容为测试用建表sql以及数据

CREATE TABLE B(
[Name] nvarchar NOT NULL,
[ID] [int] NOT NULL,
[Value] [int] NOT NULL ,
[KEY] [int] NULL,
CONSTRAINT [PK_BAS_B] PRIMARY KEY CLUSTERED
(
[Name] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]

INSERT INTO B ([Name] ,[ID] ,[Value] ,[KEY]) VALUES('张三' ,1 ,27 ,1)
INSERT INTO B ([Name] ,[ID] ,[Value] ,[KEY]) VALUES('张三' ,2 ,37 ,2)
INSERT INTO B ([Name] ,[ID] ,[Value] ,[KEY]) VALUES('李四' ,1 ,20 ,3)
INSERT INTO B ([Name] ,[ID] ,[Value] ,[KEY]) VALUES('李四' ,2 ,40 ,4)

CREATE TABLE A(
[ID] [int] NOT NULL,
[Value] [int] NOT NULL ,
[TYPE] [int] NULL,
[NAME] nvarchar NULL,
CONSTRAINT [PK_BAS_A] PRIMARY KEY CLUSTERED
(
[ID] ASC,
[Value] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]

INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,5,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,10,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,15,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,20,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,25,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,30,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,35,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,40,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,45,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (1,50,1,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,5,2,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,10,2,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,15,2,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,20,2,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,25,2,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,30,2,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,35,2,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,40,2,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,45,2,'test')
INSERT INTO A([ID],[Value],[TYPE],[NAME]) VALUES (2,50,2,'test')

刚开始之时很好奇，为什么楼主要用100C来写一个sql语句？很不明白

结果就抱着一试的态度，开始弄一下，结果就开始跟这个sql语句较劲了，我还就不信我弄不出来，这将是我写的最长的sql语句啊

花了我一个多小时，看来我的mysql功底还不行啊

1.首先根据楼主的数据库，我进行提取了关键字段，然后在本地数据库进行建表，表的结构和数据如下

2.sql语句

SELECT 
    A.id, A.value, B.key, B.id, B.value, B.type
FROM
    csdn.B B
        LEFT JOIN
    csdn.A A ON A.id = B.id
        # 对A中value和B中的value进行塞选

        # 这里需要有来两个条件来确定一条数据

        # 首先获取到根据B中value确定一个A中最大值中的最小值

        # 如果这时候单纯只根据这一个value将会有两条数据，本来应该只有一条的，因为单纯只根据 A.id = B.id和A.value = ? 存在两条数据符合

        # 这时候我们应该在根据B.value来确定这个 这就是为什么 in 前面有(A.value , B.value)
        AND (
        # --根据B.value小于A中的value的最小值-- start.....
        (A.value , B.value) IN ((SELECT 
            MIN(valueA), valueB
        FROM
            # 根据B.value小于A中的value来获取所有的A中的列
            (SELECT 
                A.value valueA, B.value valueB
            FROM
                csdn.A A, csdn.B B
            WHERE
                B.type IN (1 , 2) AND A.id = B.id
                    AND A.value > B.value) gro
        # 根据 valueB分组并获取最小的A.value
        GROUP BY valueB))

         # --根据B.value小于A中的value的最小值--  end.....

        # 根据B.value大于A中的value的最大值，和上面的情况一样
        OR (A.value , B.value) IN ((SELECT 
            MAX(valueA), valueB
        FROM
            (SELECT 
                A.value valueA, B.value valueB
            FROM
                csdn.A A, csdn.B B
            WHERE
                B.type IN (1 , 2) AND A.id = B.id
                    AND A.value < B.value) gro
        GROUP BY valueB)))

    亲测，是正确的，截图如下

    ![图片说明](https://img-ask.csdn.net/upload/201709/29/1506668917_585464.png)

select * from a left join b on a.id=b.aid

不考虑性能的情况下可以这样做，我这里提供一个思路
select A.ID,A.VALUE,A.TYPE,A.NAME,B.KEY,B.ID,B.VALUE,B.NAME
min(VALUE),max(VALUE) from A left join B on A.ID = B.ID
where B.NAME="你输入的名字"
and B.VALUE > (select min(VALUE) from B where NAME="你输入的名字"))
and B.VALUE < (select min(VALUE) from B where NAME="你输入的名字"))；
希望能帮到您

最后面的min改成max cv大法走火入魔了

select A.ID,A.VALUE,A.TYPE,A.NAME,B.KEY,B.ID,B.VALUE,B.NAME
min(VALUE),max(VALUE) from A left join B on A.ID = B.ID
where B.NAME="你输入的名字"
或者

select A.ID,A.VALUE,A.TYPE,A.NAME,B.KEY,B.ID,B.VALUE,B.NAME
min(VALUE),max(VALUE) from A left join B on A.ID = B.ID
where A.ID=''

select a.id,b.VALUE,a.type,a.name,b.key,b.id,b.value,b.name
from tb_a a
left join tb_b b
on a.id = b.id
where abs(a.value-b.value) in (
select min(abs(a.value-b.value)) tn
from tb_a a
left join tb_b b
on a.id = b.id
group by b.VALUE, b.NAME)

补一张图