许多停止和启动的设备的许多日期之间的日期差（工作时间）

Question

我有这样的桌子：

        date        | machine | activity_type | Status
31.12.2018 23:15:00 | device2 |     type2     | Start
1.1.2019   00:05:00 | device1 |     type1     | Start
1.1.2019   00:10:00 | device2 |     type2     |  Stop
1.1.2019   00:15:00 | device2 |     type1     | Start
1.1.2019   00:20:00 | device1 |     type1     |  Stop
1.1.2019   00:25:00 | device1 |     type3     | Start
2.1.2019   00:23:00 | device1 |     type3     |  Stop
3.1.2019   00:00:00 | device2 |     type1     |  Stop

我想要的输出（对于1.1.2019）将是：

date              | Duration | Activity_type | Machine
1.1.2019 00:05:00 | 00:15:00 |    type1      | device1
1.1.2019 00:00:00 | 00:10:00 |    type2      | device2
1.1.2019 00:15:00 | 23:45:00 |    type1      | device2
1.1.2019 00:25:00 | 23:35:00 |    type3      | device1

这样：

Device1从00:05:00开始，在00:20:00停止，因此持续时间为00:15:00。

Device2在00:10:00停止（它的工作从前一天开始，所以持续时间为00:10:00。

Device2从00:15:00开始，并且在这一天不会停止工作，因此持续时间为23:45:00。

Device1从00:25:00开始，并且在这一天也不会停止工作，因此持续时间为23:35:00

我想到了这个，但它一直显示NULL

SELECT
w1.Machine,
w1.Date,
w1.Type,
DateDiff(SECOND,[prevdate],[Date]) AS Duration
FROM
(
    SELECT
        Maszyna,
        Typ,
        Date,
        (
            SELECT Max(Date)
            FROM WorkHours as T2
            WHERE
                T2.Machine=T1.Machine
            AND T2.Type=T1.Type
            AND T2.Date < T1.Date
        ) AS prevdate
    FROM WorkHours AS T1
    ) as w1;

Answer 1

在这里，您已经完成了T-SQL的处理。

Schema：

CREATE TABLE Your_Table(date DATETIME,machine VARCHAR(100),activity_type VARCHAR(100), Status VARCHAR(10))

INSERT INTO Your_Table VALUES('2018-12-31 23:15:00','device2','type2','Start')
INSERT INTO Your_Table VALUES('2019-01-01 00:05:00','device1','type1','Start')
INSERT INTO Your_Table VALUES('2019-01-01 00:10:00','device2','type2','Stop')
INSERT INTO Your_Table VALUES('2019-01-01 00:15:00','device2','type1','Start')
INSERT INTO Your_Table VALUES('2019-01-01 00:20:00','device1','type1','Stop')
INSERT INTO Your_Table VALUES('2019-01-01 00:25:00','device1','type3','Start')
INSERT INTO Your_Table VALUES('2019-01-02 00:23:00','device1','type3','Stop')
INSERT INTO Your_Table VALUES('2019-01-03 00:00:00','device2','type1','Stop')
INSERT INTO Your_Table VALUES('2019-01-03 00:05:00','device2','type2','Start')
INSERT INTO Your_Table VALUES('2019-01-03 00:25:00','device2','type2','Stop')

并且我将date value传递给变量，并在查询中使用它们。

DECLARE @Date DATETIME = '2019-01-01 00:00:00',@EndDate DATETIME = '2019-01-02 00:00:00'

SELECT MIN(date),
    CONVERT(varchar(12),DATEADD(minute,DATEDIFF(MINUTE,CASE WHEN MAX(CASE WHEN Status = 'Start' THEN date END)<@Date THEN @Date ELSE MAX(CASE WHEN Status = 'Start' THEN date END) END,
             CASE WHEN MAX(CASE WHEN Status = 'Stop' THEN date END)>@EndDate THEN @EndDate ELSE MAX(CASE WHEN Status = 'Stop' THEN date END) END),0), 114)
             ,Activity_type
             ,Machine
FROM(
    SELECT *,
       ((ROW_NUMBER() OVER(PARTITION BY machine,activity_type ORDER BY date))+1)/2 RN 
    FROM Your_Table
    )A
GROUP BY Activity_type,Machine,RN

您可以在SQL Fiddle中检查输出

Answer 2

这里是一个解决方案，您将表连接到自身，并使用窗口函数为每个开始事件返回下一个相应的停止事件

查询返回从1月1日开始的所有事件及其各自的停止事件（停止事件具有相同的计算机和活动类型）

NB-我认为您期望的输出中的第二行是错误的，因为设备2的类型2没有停止事件（看起来也有错字-1.1.2019 00:00:00应该为1.1.2019 00:10:00吗？]

CREATE TABLE StartTimes
(
    [Date]          DATETIME,
    Machine         VARCHAR(7),
    Activity_Type   VARCHAR(5),
    [Status]        VARCHAR(5)
)

INSERT INTO StartTimes VALUES

('12/31/2018 23:15:00','device2','type2','Start'),
('1/1/2019   00:05:00','device1','type1','Start'),
('1/1/2019   00:10:00','device2','type2','Stop'),
('1/1/2019   00:15:00','device2','type1','Start'),
('1/1/2019   00:20:00','device1','type1','Stop'),
('1/1/2019   00:25:00','device1','type3','Start'),
('1/2/2019   00:23:00','device1','type3','Stop'),
('1/3/2019   00:00:00','device2','type1','Stop')

SELECT  StartDate,
        EndDate,
        CONVERT(varchar(12), DATEADD(MINUTE,DATEDIFF(MINUTE,StartDate,EndDate),0),114) AS Duration,
        Activity_Type,
        Machine
FROM    (
          SELECT    s.[Date] AS StartDate,
                    s.Machine,
                    s.Activity_Type,
                    s.[Status],
                    s1.[Date] AS EndDate,
                    ROW_NUMBER() OVER (PARTITION BY s.date ORDER BY s1.Date) as RowN
          FROM      StartTimes s
                    JOIN StartTimes s1
                        ON s.[Date] < s1.[Date] AND
                           s.Machine = s1.Machine AND
                           s.Activity_Type = s1.Activity_Type AND
                           s.Status = 'Start' AND
                           s1.Status = 'Stop'
       ) a
WHERE  RowN= 1 AND
       [StartDate] >= '2019-01-01' AND
       [StartDate]  < '2019-01-02'
ORDER BY StartDate