Question

在我的一个db表中，有一列FinalDate将存储日期，数据类型不是datetime而是varchar。我想写一个查询，我可以选择不同的FinalDate和分组/显示，如Jun 2012，Jul 2012。

FinalDate列的值如下所示：

那么，我如何编写一个查询来选择FinalDate的不同内容并将其显示在：

Feb 2012  
Mar 2012

Answer 1

Declare @a table (d varchar(8))
insert into @a Values ('20120213'),('20120225'),('20120218'),('20120306'),('20120320');


Select FinalDate
from
(
select Distinct 
 --DateName(Month,d)+' '+CAST(Datepart(yy,d) as Varchar(4)) as FinalDate
   SubString(DateName(Month,d),1,3)+' '+CAST(Datepart(yy,d) as Varchar(4)) as FinalDate
,Datepart(yy,d) as yy,Datepart(mm,d) as mm
from
(Select  CAST(d as datetime) as d from @a) a
) b
Order by yy,mm

Answer 2

试试这个 -

<强>查询：

DECLARE @temp TABLE (t VARCHAR(8))
INSERT INTO @temp VALUES 
    ('20120213'),
    ('20120225'),
    ('20120218'),
    ('20120306'),
    ('20120320')

SELECT LEFT(DATENAME(MONTH, t), 3) + ' ' + y
FROM (
    SELECT DISTINCT 
          t = CAST(LEFT(t, 6) + '01' AS DATETIME)
        , y = LEFT(t, 4)
    FROM @temp
) t
ORDER BY t

<强>输出：

Feb 2012
Mar 2012

Answer 3

尝试以下查询

SELECT 
Count (SUBSTRING(CONVERT(char(12), CONVERT(date, StringDate,112),113),4,12)) as counts,
 SUBSTRING(CONVERT(char(12), CONVERT(date, StringDate,112),113),4,12) from Employees
 group by SUBSTRING(CONVERT(char(12), CONVERT(date, StringDate,112),113),4,12)

输出

counts  (No column name)
1   Apr 1992 
1   Aug 1992 
1   Jan 1994 
1   Mar 1994 
1   May 1992 
1   May 1993 
1   Nov 1994 
2   Oct 1993

Answer 4

这是基于CTE的方法。将[[[TABLE_NAME]]]替换为包含FinalDate字段的表的实际名称。我在最终选择中而不是在CTE内部进行了处理，以防止SQL Server在过滤掉的（非seq 1）行上执行处理的任何可能性。

WITH a AS (
    SELECT ROW_NUMBER() OVER (PARTITION BY LEFT(FinalDate, 6) ORDER BY FinalDate) seq, FinalDate 
    FROM [[[TABLE_NAME]]]
) SELECT LEFT(DATENAME(mm, FinalDate), 3) + ' ' + LEFT(FinalDate, 4)
FROM a
WHERE seq = 1

以varchar格式选择不同日期的不同，并仅在月份和年份中显示/分组

4 个答案: