查询优化：时间间隔

总的来说，我有两种时间间隔：

presence time 和 absence time

absence time 可以具有不同的类型（例如休息，缺席，特殊日子等），并且时间间隔可能重叠和/或相交。

这是不肯定的，只有间隔的合理组合，原始数据存在，例如。重叠的存在间隔没有意义，但可能存在。我现在尝试通过多种方法来确定出现的时间间隔-对我来说，最舒服的似乎是紧随其后的时间间隔。

;with "timestamps"
as
(
    select
        "id" = row_number() over ( order by "empId", "timestamp", "opening", "type" )
        , "empId"
        , "timestamp"
        , "type"
        , "opening"
    from
    (
        select "empId", "timestamp", "type", case when "types" = 'starttime' then 1 else -1 end as "opening" from
        ( select "empId", "starttime", "endtime", 1 as "type" from "worktime" ) as data
        unpivot ( "timestamp" for "types" in ( "starttime", "endtime" ) ) as pvt
        union all
        select "empId", "timestamp", "type", case when "types" = 'starttime' then 1 else -1 end as "opening" from
        ( select "empId", "starttime", "endtime", 2 as "type" from "break" ) as data
        unpivot ( "timestamp" for "types" in ( "starttime", "endtime" ) ) as pvt
        union all
        select "empId", "timestamp", "type", case when "types" = 'starttime' then 1 else -1 end as "opening" from
        ( select "empId", "starttime", "endtime", 3 as "type" from "absence" ) as data
        unpivot ( "timestamp" for "types" in ( "starttime", "endtime" ) ) as pvt
    ) as data
)
select 
      T1."empId"
    , "starttime"   = T1."timestamp"
    , "endtime"     = T2."timestamp"
from 
    "timestamps" as T1
    left join "timestamps" as T2
        on T2."empId" = T1."empId"
        and T2."id" = T1."id" + 1
    left join "timestamps" as RS
        on RS."empId" = T2."empId"
        and RS."id" <= T1."id"      
group by
    T1."empId", T1."timestamp", T2."timestamp"
having
    (sum( power( 2, RS."type" ) * RS."opening" ) = 2)
order by 
    T1."empId", T1."timestamp";

有关一些演示数据，请参见SQL-Fiddle。

原始数据以"starttime" - "endtime"或形式存在于不同的表中"starttime" - "duration"。

想法是获得每个时间戳的有序列表，并在每个时间使用打开间隔的“位掩码”滚动总和来估计存在时间。

即使不同时间间隔的星际相等，小提琴也会起作用并给出估计的结果。在此示例中不使用索引。

这是完成质疑任务的正确方法，还是有更优雅的方法呢？

如果与回答相关：每位员工每张表格的数据量最多为一万个数据集。sql-2012无法用于总计计算内联的前辈的滚动总和。

编辑：

只需对大量测试数据（1000、10.000、100.000、100万）执行查询，就可以看到运行时间呈指数增长。显然是警告标志，对吗？

我更改了查询，并通过新奇的更新删除了滚动汇总。

我添加了一个辅助表：

create table timestamps
(
  "id" int
  , "empId" int
  , "timestamp" datetime
  , "type" int
  , "opening" int
  , "rolSum" int
)

create nonclustered index "idx" on "timestamps" ( "rolSum" ) include ( "id", "empId", "timestamp" )

我将计算滚动总和移到了这个地方：

declare @rolSum int = 0
update "timestamps" set @rolSum = "rolSum" = @rolSum + power( 2, "type" ) * "opening" from "timestamps"

看到SQL提琴这里

关于“工作时间”表中的100万个条目，运行时间减少到3秒。

问题保持不变：解决此问题的最有效方法是什么？

关于绝对最佳方法，我无法回答您的问题。但是我可以提供另一种解决问题的方法，可能更好，也可能更好。它有一个合理的执行计划，我认为它会表现良好。（我很想知道，所以分享结果！）

对于使用我自己的语法样式而不是您的语法样式，我深表歉意—当所有内容都排在通常位置时，它有助于查询向导。

该查询在SqlFiddle中可用。我为EmpID 1设置了一个重叠部分，以确保已覆盖。如果最终发现在状态数据中不会发生重叠，则可以删除最终查询和Dense_Rank计算。

WITH Points AS (
  SELECT DISTINCT
    T.EmpID,
    P.TimePoint
  FROM
    (
      SELECT * FROM dbo.WorkTime
      UNION SELECT * FROM dbo.BreakTime
      UNION SELECT * FROM dbo.Absence
    ) T
    CROSS APPLY (VALUES (StartTime), (EndTime)) P (TimePoint)
), Groups AS (
  SELECT
    P.EmpID,
    P.TimePoint,
    Grp =
      Row_Number()
      OVER (PARTITION BY P.EmpID ORDER BY P.TimePoint, X.Which) / 2
  FROM
    Points P
    CROSS JOIN (VALUES (1), (2)) X (Which)
), Ranges AS (
  SELECT
    G.EmpID,
    StartTime = Min(G.TimePoint),
    EndTime = Max(G.TimePoint)
  FROM Groups G
  GROUP BY
    G.EmpID,
    G.Grp
  HAVING Count(*) = 2
), Presences AS (
  SELECT
    R.*,
    P.Present,
    Grp =
       Dense_Rank() OVER (PARTITION BY R.EmpID ORDER BY R.StartTime)
       - Dense_Rank() OVER (PARTITION BY R.EmpID, P.Present ORDER BY R.StartTime)
  FROM
    Ranges R
    CROSS APPLY (
      SELECT
        CASE WHEN EXISTS (
          SELECT *
          FROM dbo.WorkTime W
          WHERE
            R.EmpID = W.EmpID
            AND R.StartTime < W.EndTime
            AND W.StartTime < R.EndTime
        ) AND NOT EXISTS (
          SELECT *
          FROM dbo.BreakTime B
          WHERE
            R.EmpID = B.EmpID
            AND R.StartTime < B.EndTime
            AND B.StartTime < R.EndTime
        ) AND NOT EXISTS (
          SELECT *
          FROM dbo.Absence A
          WHERE
            R.EmpID = A.EmpID
            AND R.StartTime < A.EndTime
            AND A.StartTime < R.EndTime
        ) THEN 1 ELSE 0 END
    ) P (Present)
)
SELECT
  EmpID,
  StartTime = Min(StartTime),
  EndTime = Max(EndTime)
FROM Presences
WHERE Present = 1
GROUP BY
  EmpID,
  Grp
ORDER BY
  EmpID,
  StartTime;

注意：将三个表合并在一起并添加一列以指示该时间是哪种时间：工作，休息或缺席，将提高该查询的性能。

为什么要问所有的CTE？因为每个人都是我需要对数据执行的操作。有一个汇总，或者我需要在窗口函数上放置WHERE条件，或者在不允许窗口函数的子句中使用它。

现在，我要开始看看是否能想到另一种策略来实现这一目标。:)

为了娱乐，我在此处包括为帮助解决问题而制作的“关系图”：

------------
   -----------------
                ---------------
                           -----------

    ---    ------   ------       ------------

----   ----      ---      -------

三组破折号（用空格分隔）按顺序表示：状态数据，状态数据和期望的结果。