比较SQL Server中同一列的数据

时间:2015-05-04 03:05:55

标签: sql sql-server

我的表格中有一个包含以下数据的列:

aaaa;1
aaaa;2
aaaa;3
bbbb;1
cccc;1
dddd;1
dddd;2

我需要在分号(;)之后选择编号最大的数据,如下所示:

aaaa;3
bbbb;1
cccc;1
dddd;2

有人能告诉我如何做到这一点吗?

4 个答案:

答案 0 :(得分:2)

这很简单:

select 
    -- Construct the string by left part + max(right part)
    LEFT([column], CHARINDEX(';', [column], 0) - 1) + ';' + 
    MAX(RIGHT([column], LEN([column]) - CHARINDEX(';', [column], 0)))
from 
    [table]
group by 
    LEFT([column], CHARINDEX(';', [column], 0) - 1) -- The left part of ';'

答案 1 :(得分:0)

使用CharIndex发现分号的位置,然后使用结果进行排序。示例代码如下:

declare @table table (
col1 varchar(25)
)

insert into @table (col1) values ('aaaa;1')
insert into @table (col1) values ('aaaa;2')
insert into @table (col1) values ('aaaa;3')
insert into @table (col1) values ('bbbb;1')
insert into @table (col1) values ('dddd;1')
insert into @table (col1) values ('dddd;2')

select top 1
col1,
charindex(';',col1,0) as SemiColonLocation,
substring(col1, 0, charindex(';',col1,0) + 1) as TextVal,
substring(col1, charindex(';',col1,0) + 1, (len(col1) - charindex(';',col1,0))) as AfterVal
from @table
order by substring(col1, charindex(';',col1,0) + 1, (len(col1) - charindex(';',col1,0))) desc

答案 2 :(得分:0)

您可以使用LEFTSUBSTRINGCHARINDEX功能来实现此目的。有关详细信息,请阅读此article

SQL Fiddle

WITH CteSplit(string, leftString, rightString, RN) AS(
    SELECT 
        string,
        LEFT(string, CHARINDEX(';', string, 0) - 1),
        CAST(SUBSTRING(string, CHARINDEX(';', string, 0) + 1, LEN(string) - CHARINDEX(';', string, 0)) AS INT),
        RN = ROW_NUMBER() OVER(
                PARTITION BY 
                    LEFT(string, CHARINDEX(';', string, 0) - 1)
                ORDER BY
                    CAST(SUBSTRING(string, CHARINDEX(';', string, 0) + 1, LEN(string) - CHARINDEX(';', string, 0)) AS INT) DESC
            )
    FROM TestTable
)
SELECT
    string
FROM CteSplit
WHERE RN = 1

答案 3 :(得分:0)

你可以像这样使用rank();


    select column_name
      from (select a.*,
                   rank() over(partition by substr(column_name, 1, 4) order by substr(column, 6) desc) as row_num
              from table_name a)
     where row_num = '1';