SQL 按相似条件分组

Question

假设有一个 SQL table: 包含列的 testTable：clientID、colA、colB、colC。

reference   clientID    colA    colB    colC
---------------------------------------------
001            1        test1   test2   test3
002            1        test1   ball2   test3
003            2        test1   ball2   test3
004            2        test1   ball2   test3
005            3        test1   test2   test3
006            4        test1   test2   test3
007            4        test1   test2   test3
009            5        test1   ball2   test3
---------------------------------------------

我想 select 所有不同的行，其中 colB 类似于 'test' 并按 clientID 分组。所以我最终得到：

reference    clientID    colA    colB    colC
----------------------------------------------
001             1        test1   test2   test3
005             3        test1   test2   test3
006             4        test1   test2   test3
----------------------------------------------

编辑：参考列是唯一的如果我使用 select distinct * .. from .. where colB like '%test%' group by clientID 那么返回的结果没有分组的clientID

Answer 1

仅仅一个不同的是不够的？

Select distinct * from #data where ColB like 'test%'

或者，如果您只需要考虑某些特定的列，则可以使用前 1 位并列

Select top (1) with ties * from #groupby where ColB like 'test%'
order by row_number() over(partition by clientid order by cola,colb,colc)--you can include your required columns only

Answer 2

where 子句位于 having 和 group by 之前。如果你想在分组前过滤掉记录条件在where子句中，如果你想过滤掉分组记录条件在having子句中：

select ...
from ...
where ...
group by ...
having ...

或者类似这样的东西，如果上面的东西不起作用的话

select ...
from (
   select ...
   from ...
   where ...
   group by ...
   having ...
) x
where ...

Answer 3

试试这个，

 Select distinct * From tableA where colB like '%test%'

Answer 4

当您使用 GROUP BY 时，您必须指定您在查询中使用且不使用聚合函数的所有列。所以，例如，

SELECT CLIENTID, COLA, COLB, COLC, COUNT(*) AS RC
FROM YOURTABLE
WHERE COLB LIKE '%test%'
GROUP BY CLIENTID, COLA, COLB, COLC

Answer 5

按一列分组时，会将多行合并为一行，select中的其他列必须是聚合函数或子查询。使用哪个功能取决于您的需要。如果与字符串 column

一起使用，如下例所示使用 MIN() 将按字母顺序给出第一个结果

SELECT clientID
    , MIN(colA) AS colA
    , MIN(colB) AS colB
    , MIN(colC) AS colC
FROM tableA
WHERE colB LIKE '%test%'
GROUP BY clientID

编辑：这是另一种解决方案，不使用 GROUP BY，而是使用 Common Table Expression using ROW_NUMBER()

WITH CTE_Source AS 
(
  SELECT *
  FROM TableA
  WHERE colB LIKE '%test%'
) 
, CTE_Filter AS 
(
   SELECT *, ROW_NUMBER() OVER (PARTITION BY ClientID ORDER BY reference) RN
   FROM CTE_Source 
)
SELECT * 
FROM CTE_Filter 
WHERE RN = 1

Answer 6

只需使用：-

group by  clientID,colA,colB,colC

而不是

group by  clientID

演示：-

Create table testTable (clientID int , colA varchar(10), colB varchar(10), colC varchar(10))
go
insert into testTable values
(1,'test1','test2','test3'),
(1,'test1','ball2','test3'),
(2,'test1','ball2','test3'),
(2,'test1','ball2','test3'),
(3,'test1','test2','test3'),
(4,'test1','test2','test3'),
(4,'test1','test2','test3'),
(5,'test1','ball2','test3')


select * from testTable
where colB LIKE '%test%'
group by  clientID,colA,colB,colC

输出：-

clientID    colA    colB    colC
  1        test1    test2   test3
  3        test1    test2   test3
  4        test1    test2   test3

SQL 按相似条件分组

SQL group by with like condition

sql

sql-server

group-by

where

sql-like