StartDate 和 EndDate 之间的季度天数 - SQL Server答案

【问题标题】：Number of days by quarter between StartDate and EndDate - SQL ServerStartDate 和 EndDate 之间的季度天数 - SQL Server
【发布时间】：2019-05-18 21:20:47
【问题描述】：

我需要按季度的天数填写列。时间周期约为 10 年。表结构如下。感谢您的任何建议

SELECT StartDate, EndDate
INTO #tmp_RTX 
FROM DateTable


         StartDate                 EndDate
-------------------------------------------------
2015-11-01 00:00:00.000  2018-06-01 00:00:00.000
2017-09-02 00:00:00.000  2021-12-02 00:00:00.000
2016-01-02 00:00:00.000  2019-01-02 00:00:00.000
      .
      .
      .

2018-10-26 00:00:00.000  2020-10-26 00:00:00.000

  INSERT INTO DWHMart.[RTX].[RoadTax]
      (  ,[NuberOfaDaysInQuarter1]
         ,[NuberOfaDaysInQuarter2]
         ,[NuberOfaDaysInQuarter3]
         ,[NuberOfaDaysInOctoberNovember]
         ,[NuberOfaDaysInDecember]
         ,[NuberOfaDaysInTotal])

SELECT 
,[NuberOfaDaysInQuarter1] = NuberOfaDaysInQuarter1
,[NuberOfaDaysInQuarter2] = NuberOfaDaysInQuarter2
,[NuberOfaDaysInQuarter3] = NuberOfaDaysInQuarter3
,[NuberOfaDaysInOctoberNovember] = NuberOfaDaysInOctoberNovember
,[NuberOfaDaysInDecember] = NuberOfaDaysInDecember
,[NuberOfaDaysInTotal] = DATEDIFF(DAY, StartDate, EndDate)
FROM #tmp_RTX

预期结果：

StartDate                 EndDate                 NuberOfaDaysInQuarter1   ..
2015-11-01 00:00:00.000  2018-06-01 00:00:00.000        .
2017-09-02 00:00:00.000  2021-12-02 00:00:00.000        .
2016-01-02 00:00:00.000  2019-01-02 00:00:00.000        .
      .                                                 .
      .                                                 .
      .                                                 .

2018-10-26 00:00:00.000  2020-10-26 00:00:00.000        .

【问题讨论】：

欢迎来到 Stack Overflow。我担心您的目标在这里相当模糊。您是在寻求建议，而不是帮助您解决遇到问题的特定代码。纯粹猜测，您似乎需要Calendar Table。否则，您将需要扩展您的问题。显示您拥有的当前数据，以及您追求的预期结果；以及您的尝试。谢谢。
您尚未定义您的季度是否遵循日历年季度，或者它们是否处于不同的会计年度变体中。
季度在哪一年也有关系吗？
关于日历年。是的，这很重要@xQbert

标签： sql-server date select time period

【解决方案1】：

我使用递归 CTE 制作了自己的动态天数表。

然后我在你给出的每个范围内按四分之一计算天数

我对您的每一行都使用了 CROSS JOIN，因为它们没有唯一的 GROUP BY ID，但无论如何它都可以工作

;WITH RNG AS (SELECT cast(MIN(StartDate) as date) as MN, 
                 cast(MAX (EndDate ) as date) MX FROM #tmp_RTX),
 DATS AS (SELECT MN FROM RNG
          UNION ALL 
          SELECT DATEADD(day,1,MN) FROM dats 
            WHERE dats.MN < (SELECT MX FROM RNG)
        )
        SELECT StartDate T1,
                Enddate T1, 
                DQ.NuberOfaDaysInQuarter1   ,
                DQ.NuberOfaDaysInQuarter2   ,
                DQ.NuberOfaDaysInQuarter3  , 
                DQ.NuberOfaDaysInQuarter4   
                    FROM #tmp_RTX T1
                        CROSS APPLY (SELECT 
                                    SUM(CASE WHEN datepart(quarter ,MN) = 1 THEN 1 ELSE 0 END) 
                                            NuberOfaDaysInQuarter1,
                                    SUM(CASE WHEN datepart(quarter ,MN) = 2 THEN 1 ELSE 0 END) 
                                            NuberOfaDaysInQuarter2,      
                                    SUM(CASE WHEN datepart(quarter ,MN) = 3 THEN 1 ELSE 0 END) 
                                            NuberOfaDaysInQuarter3,
                                    SUM(CASE WHEN datepart(quarter ,MN) = 4 THEN 1 ELSE 0 END) 
                                            NuberOfaDaysInQuarter4
                                                 FROM DATS
                                                 WHERE MN BETWEEN 
                                                            CAST(t1.startdate as date) 
                                                                and 
                                                                CAST(T1.EndDate as date)
                            ) DQ

        OPTION  (maxrecursion 0);

以下是构建我的测试数据的内容（基于您的数据约定）

SELECT '20151101' StartDate, '20180601' EndDate
INTO #tmp_RTX 

INSERT into #tmp_RTX  values ('20161231','20180101');

INSERT into #tmp_RTX  values ('20181231','20190101');

要处理最后一个季度的不同要求，请检查月份，而不是季度

;WITH RNG AS (SELECT cast(MIN(StartDate) as date) as MN, 
                 cast(MAX (EndDate ) as date) MX FROM #tmp_RTX),
 DATS AS (SELECT MN FROM RNG
          UNION ALL 
          SELECT DATEADD(day,1,MN) FROM dats 
            WHERE dats.MN < (SELECT MX FROM RNG)
        )
        SELECT StartDate T1,
                Enddate T1, 
                DQ.NuberOfaDaysInQuarter1   ,
                DQ.NuberOfaDaysInQuarter2   ,
                DQ.NuberOfaDaysInQuarter3  , 
                DQ.NuberOfaDaysInOctoberNovember   ,
                DQ.NuberOfaDaysInDecember
                    FROM #tmp_RTX T1
                        CROSS APPLY (SELECT 
                                    SUM(CASE WHEN datepart(quarter ,MN) = 1 THEN 1 ELSE 0 END) 
                                            NuberOfaDaysInQuarter1,
                                    SUM(CASE WHEN datepart(quarter ,MN) = 2 THEN 1 ELSE 0 END) 
                                            NuberOfaDaysInQuarter2,      
                                    SUM(CASE WHEN datepart(quarter ,MN) = 3 THEN 1 ELSE 0 END) 
                                            NuberOfaDaysInQuarter3,
                                    SUM(CASE WHEN datepart(month ,MN) IN(10,11) THEN 1 ELSE 0 END) 
                                            NuberOfaDaysInOctoberNovember,
                                    SUM(CASE WHEN datepart(month ,MN) IN(12) THEN 1 ELSE 0 END) 
                                            NuberOfaDaysInDecember
                                                 FROM DATS
                                                 WHERE MN BETWEEN 
                                                            CAST(t1.startdate as date) 
                                                                and 
                                                                CAST(T1.EndDate as date)
                            ) DQ

        OPTION  (maxrecursion 0)

【讨论】：

干得好，但是您将如何为 [NuberOfaDaysInOctoberNovember] 和 [NuberOfaDaysInDecember] 而不是 NuberOfaDaysInQuarter4 做到这一点？

【解决方案2】：

除闰年外，季度的天数始终相同。因此，您可以使用 Q2..Q4 的常量表达式和 Q1 的参数化值。

DECLARE @my_date date = '2017-01-01';
SELECT 89 + ISDATE(CAST(@my_date AS char(4)) + '0229') AS NuberOfaDaysInQuarter1,
       90 AS NuberOfaDaysInQuarter2,
       91 AS NuberOfaDaysInQuarter3,
       91 AS NuberOfaDaysInQuarter4

更新的解决方案（按季度累积天数）

DECLARE @start_date date = '2000-01-01';
DECLARE @quarters TABLE (start_date date, end_date date, num int);
-- Create quarter calendar from @start_date
WITH nums (n) AS
(
   SELECT TOP 100 (ROW_NUMBER() OVER (ORDER BY (SELECT 1)) - 1) * 3
   FROM sys.columns sc1 CROSS JOIN sys.columns sc2
),
calendar AS
(
   SELECT n, DATEADD(month, n, @start_date) AS start_date
   FROM nums
)
INSERT INTO @quarters
SELECT start_date, 
       DATEADD(day, -1, DATEADD(month, 3, start_date)) AS end_date,
       ((n / 3) % 4) + 1 AS num
FROM calendar;

-- test data
DECLARE @t TABLE (start_date date, end_date date);
INSERT INTO @t VALUES
('2015-11-01', '2018-06-01'),
('2017-09-02', '2021-12-02'),
('2016-01-02', '2019-01-02');

-- Calculations
WITH inner_counts AS (
   SELECT t.start_date, t.end_date, q.num,
          SUM(datediff(day, q.start_date, q.end_date)) AS days_count
   FROM @t t INNER JOIN @quarters q ON q.start_date > t.start_date AND q.start_date < t.end_date
   GROUP BY t.start_date, t.end_date, q.num
),
outer_counts1 AS (
   SELECT t.start_date, t.end_date, q.num,
          SUM(datediff(day, t.start_date, q.end_date)) AS days_count
   FROM @t t INNER JOIN @quarters q ON q.end_date = (SELECT MIN(end_date) FROM @quarters q WHERE q.end_date >= t.start_date)
   GROUP BY t.start_date, t.end_date, q.num
),
outer_counts2 AS (
   SELECT t.start_date, t.end_date, q.num,
          SUM(datediff(day, q.start_date, t.end_date)) AS days_count
   FROM @t t INNER JOIN @quarters q ON q.start_date = (SELECT MAX(start_date) FROM @quarters q WHERE q.start_date <= t.end_date)
   GROUP BY t.start_date, t.end_date, q.num
),
total_counts AS (
   SELECT start_date, end_date, num, SUM(days_count) AS days_count
   FROM (SELECT * FROM inner_counts 
         UNION ALL 
         SELECT * FROM outer_counts1 
         UNION ALL 
         SELECT * FROM outer_counts2) c
   GROUP BY start_date, end_date, num
)
SELECT start_date, end_date,
       SUM(CASE WHEN num = 1 THEN days_count ELSE 0 END) AS days_in_Q1,
       SUM(CASE WHEN num = 2 THEN days_count ELSE 0 END) AS days_in_Q2,
       SUM(CASE WHEN num = 3 THEN days_count ELSE 0 END) AS days_in_Q3,
       SUM(CASE WHEN num = 4 THEN days_count ELSE 0 END) AS days_in_Q4
FROM total_counts
GROUP BY start_date, end_date

结果

start_date end_date   days_in_Q1  days_in_Q2  days_in_Q3  days_in_Q4
---------- ---------- ----------- ----------- ----------- -----------
2015-11-01 2018-06-01 268         331         182         242
2016-01-02 2019-01-02 357         270         273         273
2017-09-02 2021-12-02 357         360         392         517

【讨论】：

但是我需要 StartDate 和 EndDate 之间的季度天数
@Michal 什么是“按季度计算的天数”？
DATEDIFF（日期，开始日期，结束日期）。我需要分成 quartals
@Michal 所以问题是“两个日期之间的季度数”？该季度的天数是恒定的。
例如：当 StartDate 为 2015-03-15 和 EndDate 2015-04-05 时，每个季度的天数不能相同。第 1 季度应为 16（天），第 2 季度应为 5（天）。

【解决方案3】：

DEMO

使用递归 CTE（请务必将最大递归选项设置为您需要循环处理的天数，或者在您确认不存在无限循环后设置为 0。）生成您的最大和最小天数之间的所有天数日期表中的日期，然后将此结果加入到日期范围之间的日期表中。在周期日期的年份和季度上使用日期部分，我们可以获得适当的季度/年度，然后进行分组。

我没有调整结果。透视数据通常是一种显示功能，最好留给那些显示工具 Crystal、BI 等。它可以在 SQL 中完成，在这种情况下，将需要动态 SQL，因为年数本质上是动态的。我选择不走这条路，因为报告工具可以处理数据的动态特性。

WITH 
DateTable as (SELECT cast('2015-11-01 00:00:00.000'as  date) startdate,  cast('2018-06-01 00:00:00.000'as  date) endDate union all
              SELECT '2017-09-02 00:00:00.000',  '2021-12-02 00:00:00.000' union all
              SELECT '2016-01-02 00:00:00.000',  '2019-01-02 00:00:00.000'),

CTE  AS (SELECT Min(StartDate) StartDate, max(EndDate) EndDate FROM DATETABLE),

CTE2 AS (SELECT C.StartDate as RangeStartDate
              , C.EndDate as RangeEndDate
              , 1 CycleCount
              , datepart(Q,dateadd(d,0,C.StartDate)) as Quarter
              , datepart(m,dateadd(d,0,C.StartDate)) as Month
              , datepart(YYYY, dateadd(d,0,C.StartDate)) as Yr
              , C.StartDate as CycleDate
         FROM CTE C
         UNION ALL
         SELECT RangeStartDate, RangeEndDate, CycleCount+1
              , datepart(Q,dateadd(d,1,CycleDate)) as Quarter
              , datepart(m,dateadd(d,1,CycleDate)) as Month     
              , datepart(YYYY, dateadd(d,1,CycleDate)) as Yr
              , dateadd(d,1,cycleDate) as CycleDate
         FROM  cte2
         WHERE datediff(d,RangeStartDate, RangeEndDate) >= CycleCount 

    )
    SELECT DT.StartDate
         , DT.ENDDate
         , concat(yr,'-',Quarter) [YYYY-Q]
         , count(*)  as DaysInQuarter
         , sum(case when Month in (10,11) then 1 else 0 end)as OctNovDays
         , sum(case when Month in (12) then 1 else 0 end)as DecDays
         , datediff(d,StartDate, EndDate) as DaysTotal
    FROM CTE2
    INNER JOIN DateTable DT
      on CTE2.CycleDate between DT.StartDate and DT.EndDate
    GROUP BY Quarter,DT.StartDate, DT.EndDate, YR
    ORDER BY  DT.StartDate, DT.EndDate, YR, Quarter
    OPTION (MAXRECURSION 10000)

给我们：

+----+---------------------+---------------------+--------+---------------+------------+---------+-----------+
|    |      StartDate      |       ENDDate       | YYYY-Q | DaysInQuarter | OctNovDays | DecDays | DaysTotal |
+----+---------------------+---------------------+--------+---------------+------------+---------+-----------+
|  1 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2015-4 |            61 |         30 |      31 |       943 |
|  2 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2016-1 |            91 |          0 |       0 |       943 |
|  3 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2016-2 |            91 |          0 |       0 |       943 |
|  4 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2016-3 |            92 |          0 |       0 |       943 |
|  5 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2016-4 |            92 |         61 |      31 |       943 |
|  6 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2017-1 |            90 |          0 |       0 |       943 |
|  7 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2017-2 |            91 |          0 |       0 |       943 |
|  8 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2017-3 |            92 |          0 |       0 |       943 |
|  9 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2017-4 |            92 |         61 |      31 |       943 |
| 10 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2018-1 |            90 |          0 |       0 |       943 |
| 11 | 01.11.2015 00:00:00 | 01.06.2018 00:00:00 | 2018-2 |            62 |          0 |       0 |       943 |
| 12 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2016-1 |            90 |          0 |       0 |      1096 |
| 13 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2016-2 |            91 |          0 |       0 |      1096 |
| 14 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2016-3 |            92 |          0 |       0 |      1096 |
| 15 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2016-4 |            92 |         61 |      31 |      1096 |
| 16 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2017-1 |            90 |          0 |       0 |      1096 |
| 17 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2017-2 |            91 |          0 |       0 |      1096 |
| 18 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2017-3 |            92 |          0 |       0 |      1096 |
| 19 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2017-4 |            92 |         61 |      31 |      1096 |
| 20 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2018-1 |            90 |          0 |       0 |      1096 |
| 21 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2018-2 |            91 |          0 |       0 |      1096 |
| 22 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2018-3 |            92 |          0 |       0 |      1096 |
| 23 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2018-4 |            92 |         61 |      31 |      1096 |
| 24 | 02.01.2016 00:00:00 | 02.01.2019 00:00:00 | 2019-1 |             2 |          0 |       0 |      1096 |
| 25 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2017-3 |            29 |          0 |       0 |      1552 |
| 26 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2017-4 |            92 |         61 |      31 |      1552 |
| 27 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2018-1 |            90 |          0 |       0 |      1552 |
| 28 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2018-2 |            91 |          0 |       0 |      1552 |
| 29 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2018-3 |            92 |          0 |       0 |      1552 |
| 30 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2018-4 |            92 |         61 |      31 |      1552 |
| 31 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2019-1 |            90 |          0 |       0 |      1552 |
| 32 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2019-2 |            91 |          0 |       0 |      1552 |
| 33 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2019-3 |            92 |          0 |       0 |      1552 |
| 34 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2019-4 |            92 |         61 |      31 |      1552 |
| 35 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2020-1 |            91 |          0 |       0 |      1552 |
| 36 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2020-2 |            91 |          0 |       0 |      1552 |
| 37 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2020-3 |            92 |          0 |       0 |      1552 |
| 38 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2020-4 |            92 |         61 |      31 |      1552 |
| 39 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2021-1 |            90 |          0 |       0 |      1552 |
| 40 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2021-2 |            91 |          0 |       0 |      1552 |
| 41 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2021-3 |            92 |          0 |       0 |      1552 |
| 42 | 02.09.2017 00:00:00 | 02.12.2021 00:00:00 | 2021-4 |            63 |         61 |       2 |      1552 |
+----+---------------------+---------------------+--------+---------------+------------+---------+-----------+

【讨论】：

已更新以处理 10 月/11 月和 12 月的列和总数；我在原始帖子中没有考虑到这一点。