select Columns1 ,min(Columns2) from tb group by Columns1

------解决方案--------------------
SQL code
select * from tb t where columns2=(select min(columns2) from tb where columns=t.columns)

------解决方案--------------------
select Columns1 ,min(Columns2) from tb group by Columns1
select * from tb t where columns2=(select min(columns2) from tb where columns=t.columns)

两个都可以，那假如最小的col2有重复的怎么处理？是全部显示出来还是只取一个？


------解决方案--------------------
可能给的数据、结果不够典型
否则3楼的sql就行了

------解决方案--------------------
select * from (
select Columns1, Columns2,row_number() over (partition by Columns1 order by Columns2) cnt from table )
where cnt =1



------解决方案--------------------
SQL code

select Columns1,min(Columns2) from tb group by Columns1

------解决方案--------------------
探讨

/*
Columns1 Columns2
A1 B1
A1 B1
A1 B2
A2 B2
A2 B3
A3 B1
*/
-- Expect result data:
/*
Columns1 Columns2
A1 B1
A2 B2
A3 B1
*/


------解决方案--------------------
SQL code

CREATE TABLE #TEMP(ID VARCHAR(2),ID2 VARCHAR(2))
INSERT INTO #TEMP
SELECT 'A1','B1'
UNION ALL
SELECT 'A1','B1'
UNION ALL
SELECT 'A1','B2'
UNION ALL
SELECT 'A2','B2'
UNION ALL
SELECT 'A2','B3'
UNION ALL
SELECT 'A3','B1'


SELECT DISTINCT * FROM #TEMP T0 WHERE NOT EXISTS(SELECT 1 FROM #TEMP WHERE T0.ID=ID AND T0.ID2>ID2)

/*
ID ID2
A1 B1
A2 B2
A3 B1
*/

------解决方案--------------------
問清他的需求.有岐意.

------解决方案--------------------
不太懂 。。。。 哎