0

我有一些数据，例如：

t   starttime                 endtime

1   2012-11-13 07:00:00.000   2012-11-13 09:00:00.000
1   2012-11-13 09:00:00.000   2012-11-13 09:15:00.000
2   2012-11-13 09:00:00.000   2012-11-13 16:00:00.000
1   2012-11-13 09:15:00.000   2012-11-13 12:00:00.000
1   2012-11-13 12:30:00.000   2012-11-13 15:00:00.000
1   2012-11-14 07:00:00.000   2012-11-14 09:00:00.000
1   2012-11-14 09:00:00.000   2012-11-14 09:15:00.000
1   2012-11-14 09:15:00.000   2012-11-14 12:00:00.000
1   2012-11-14 12:30:00.000   2012-11-14 15:00:00.000
1   2012-11-15 07:00:00.000   2012-11-15 09:00:00.000
1   2012-11-15 09:00:00.000   2012-11-15 09:15:00.000
1   2012-11-15 09:15:00.000   2012-11-15 12:00:00.000
1   2012-11-15 12:30:00.000   2012-11-15 15:00:00.000

并希望对其进行预处理，例如：

timestamp               t   x

2012-11-13 07:00:00.000 1   1
2012-11-13 09:00:00.000 1   0
2012-11-13 09:00:00.000 1   1
2012-11-13 09:00:00.000 2   1
2012-11-13 09:15:00.000 1   0
2012-11-13 09:15:00.000 1   1
2012-11-13 12:00:00.000 1   0
2012-11-13 12:30:00.000 1   1
2012-11-13 15:00:00.000 1   0
2012-11-13 16:00:00.000 2   0
2012-11-14 07:00:00.000 1   1
2012-11-14 09:00:00.000 1   0
2012-11-14 09:00:00.000 1   1
2012-11-14 09:15:00.000 1   0
2012-11-14 09:15:00.000 1   1
2012-11-14 12:00:00.000 1   0
2012-11-14 12:30:00.000 1   1
2012-11-14 15:00:00.000 1   0
2012-11-15 07:00:00.000 1   1
2012-11-15 09:00:00.000 1   0
2012-11-15 09:00:00.000 1   1
2012-11-15 09:15:00.000 1   0
2012-11-15 09:15:00.000 1   1
2012-11-15 12:00:00.000 1   0
2012-11-15 12:30:00.000 1   1
2012-11-15 15:00:00.000 1   0

x应该定义，如果值timestamp来自starttime(x = 1) 或来自endtime(x = 0)。我知道，这是小菜一碟，使用两个selects和一个union all。

但我想知道是否可以使用unpivot()一些扩展？开始像：

SELECT 
    , UD."timeStamp"
    , UD."t"
    , "isStart"
FROM 
    (
        SELECT 
              "StartTime"
            , "EndTime"
            , "t"
            , < REFERENCE TO STARTTIME OR ENDTIME ? >
        FROM 
            Data 
    ) AS BC
UNPIVOT ( "timeStamp" FOR dummy IN ( "StartTime", "EndTime" )) AS UD

< REFERENCE TO STARTTIME OR ENDTIME ? >应该提供信息，从哪个列 timeStamp 来自...

4

1 回答 1

2

像这样的东西怎么样，只需CASE对其应用一个声明：

select timestamp,
  t,
  case when col = 'starttime' then 1 else 0 end x
from yourtable
unpivot
(
  timestamp for col in (starttime, endtime)
) un

请参阅带有演示的 SQL Fiddle

如果您不使用该CASE语句，则该col字段仅显示值starttime或endtime.

结果：

|                       TIMESTAMP | T | X |
-------------------------------------------
| November, 13 2012 07:00:00+0000 | 1 | 1 |
| November, 13 2012 09:00:00+0000 | 1 | 0 |
| November, 13 2012 09:00:00+0000 | 1 | 1 |
| November, 13 2012 09:15:00+0000 | 1 | 0 |
| November, 13 2012 09:00:00+0000 | 2 | 1 |
| November, 13 2012 16:00:00+0000 | 2 | 0 |
| November, 13 2012 09:15:00+0000 | 1 | 1 |
| November, 13 2012 12:00:00+0000 | 1 | 0 |
| November, 13 2012 12:30:00+0000 | 1 | 1 |
| November, 13 2012 15:00:00+0000 | 1 | 0 |
| November, 14 2012 07:00:00+0000 | 1 | 1 |
| November, 14 2012 09:00:00+0000 | 1 | 0 |
| November, 14 2012 09:00:00+0000 | 1 | 1 |
| November, 14 2012 09:15:00+0000 | 1 | 0 |
| November, 14 2012 09:15:00+0000 | 1 | 1 |
| November, 14 2012 12:00:00+0000 | 1 | 0 |
| November, 14 2012 12:30:00+0000 | 1 | 1 |
| November, 14 2012 15:00:00+0000 | 1 | 0 |
| November, 15 2012 07:00:00+0000 | 1 | 1 |
| November, 15 2012 09:00:00+0000 | 1 | 0 |
| November, 15 2012 09:00:00+0000 | 1 | 1 |
| November, 15 2012 09:15:00+0000 | 1 | 0 |
| November, 15 2012 09:15:00+0000 | 1 | 1 |
| November, 15 2012 12:00:00+0000 | 1 | 0 |
| November, 15 2012 12:30:00+0000 | 1 | 1 |
| November, 15 2012 15:00:00+0000 | 1 | 0 |
于 2012-11-14T10:50:39.933 回答