我很难弄清楚这一点请帮助我 我有时间进/出sql查询 我有一张桌子如下图所示。有四列显示时间输入/输出信息，例如......
Date Day Day TimeStamp CheckType
10/11/2014 Sat 8:30 am in
10/11/2014 Sat 11:30am out
10/11/2014 Sat 1:30pm in
10/11/2014 Sat out
10/12/2014 Sun 9:00am in
10/12/2014 Sun 11:20pm out
10/12/2014 Sun 5:20pm out
10/13/2014 Mon 8:00am in
10/13/2014 Mon 6:10pm in
所以无论谁签入或结账，然后记录将按顺序显示结果，如果有人应该退房，但意外按下按钮，那么这将显示为（in）或者如果有人忘记退房那么空间将显示空白
我正在尝试将行转换为列并在下面显示此类信息
Date Day Time Type Time Type Time Type Time Type etc-----
10/11/2014 Sat 8:30am in 11:30am out 1:30pm in
10/12/2014 Sun 9:00am in 11:20am out 1:20pm in 6:20pm in
10/13/2014 Mon 8:00am in 6:10pm out
10/14/2014 Tus 8:20am in
etc
我尝试使用pivot
select Date, Day, [1],[2],[3],[4],[5],[6],[7],[8],[9],[10] etc---
from
(
select Date, Day, Stamptime, CheckTime, userID
from a table
)
pivot
(
max(StampTime)
for stamptime in ([1],[2],[3],[4],[5],[6],[7],[8],[9],[10] etc---)
) as PivotTable
任何人都可以解释如何将行转换为我已经花了很多天的列。
答案 0 :(得分：2)
这里有一些与你所追求的相近的东西，利用XML来获得可变数量的列。正如我在上面的评论中所提到的，我不推荐这种方法。 SQL小提琴：http://sqlfiddle.com/#!3/e5b325/2
select [Date]
, [Day]
, (
select [TimeStamp] [Time]
, [CheckType] [Type]
from aTable b
where b.[Date] = a.[Date]
order by [TimeStamp], [CheckType]
for xml path ('')
) CheckInAndOutInfo
from aTable a
group by [Date], [Day]
order by [Date]
输出：
DATE DAY CHECKINANDOUTINFO
2014-10-11 Sat <Type>Out</Type><Time>08:30:00</Time><Type>In</Type><Time>11:30:00</Time><Type>Out</Type><Time>13:30:00</Time><Type>In</Type>
2014-10-12 Sun <Time>09:00:00</Time><Type>In</Type><Time>17:20:00</Time><Type>Out</Type><Time>23:20:00</Time><Type>Out</Type>
2014-10-13 Mon <Time>08:00:00</Time><Type>In</Type><Time>18:10:00</Time><Type>In</Type>
或者，如果您可以保证每天的登记/退出数量不会超过一定数量，您可以执行以下操作（假设每天不超过5次）： SQL小提琴：http://sqlfiddle.com/#!3/e5b325/4
select *
from
(
select [Date], [Day]
, 'T' + CAST(ROW_NUMBER() over (partition by [Date] order by [TimeStamp], [CheckType]) as nvarchar) r
, cast([TimeStamp] as nvarchar) pvtVal
from aTable
where [TimeStamp] is not null
union all
select [Date], [Day]
, 'C' + CAST(ROW_NUMBER() over (partition by [Date] order by [TimeStamp], [CheckType]) as nvarchar) r
, cast([CheckType] as nvarchar) pvtVal
from aTable
where [TimeStamp] is not null
) x
pivot
(
min(pvtVal)
for r in ([T1], [C1], [T2], [C2], [T3], [C3], [T4], [C4], [T5], [C5])
) y
order by [Date]
输出：
DATE DAY T1 C1 T2 C2 T3 C3 T4 C4 T5 C5
2014-10-11 Sat 08:30:00.0000000 In 11:30:00.0000000 Out 13:30:00.0000000 In (null) (null) (null) (null)
2014-10-12 Sun 09:00:00.0000000 In 17:20:00.0000000 Out 23:20:00.0000000 Out (null) (null) (null) (null)
2014-10-13 Mon 08:00:00.0000000 In 18:10:00.0000000 In (null) (null) (null) (null) (null) (null)
...或者如果你想使用动态SQL，你可以这样做： SQL小提琴：http://sqlfiddle.com/#!3/e5b325/6
declare @sql nvarchar(max)
select @sql = coalesce(@sql+',','') + QUOTENAME('T' + CAST(x as nvarchar)) + ',' + QUOTENAME('C' + CAST(x as nvarchar))
from
(
select distinct row_number() over (partition by [Date] order by [Date]) x
from aTable
where [TimeStamp] is not null
) y
order by x
set @sql =
'select *
from
(
select [Date], [Day]
, ''T'' + CAST(ROW_NUMBER() over (partition by [Date] order by [TimeStamp], [CheckType]) as nvarchar) r
, cast([TimeStamp] as nvarchar) pvtVal
from aTable
where [TimeStamp] is not null
union all
select [Date], [Day]
, ''C'' + CAST(ROW_NUMBER() over (partition by [Date] order by [TimeStamp], [CheckType]) as nvarchar) r
, cast([CheckType] as nvarchar) pvtVal
from aTable
where [TimeStamp] is not null
) x
pivot
(
min(pvtVal)
for r in (' + @sql + ')
) y
order by [Date]'
exec (@sql)
输出：
DATE DAY T1 C1 T2 C2 T3 C3
2014-10-11 Sat 08:30:00.0000000 In 11:30:00.0000000 Out 13:30:00.0000000 In
2014-10-12 Sun 09:00:00.0000000 In 17:20:00.0000000 Out 23:20:00.0000000 Out
2014-10-13 Mon 08:00:00.0000000 In 18:10:00.0000000 In (null) (null)