在 SQL Server 中的 2 个日期之间添加缺少的开始日期的 SQL 代码

Question

我有按 clientID、contractID 和 effectiveDate 排序的下表。一个客户有多个 contractID 和它各自的 effectiveDate。

所需的输出如下，其中新的 FYStartDate 列应在 clientID 的后续 contractID 的 2 个日期之间添加缺失的 FYStartDate（在这种情况下，财政年度从每年的 01June 开始）

如果您能分享所需的 SQL 代码，我将不胜感激。

我附上了生成第一个表的 SQL 代码

CREATE TABLE [client] (
[clientid] [int] NULL,
[contractid] [int] NULL,
[effectivedate] [date] NULL
) ON [PRIMARY]
GO 

insert into [client] values
('228','2','6/1/2003'),('228','136','6/1/2004'),('228','242','6/1/2008'), 
('228','337','12/1/2012'),('228','584','6/1/2017'),('14216','319','5/1/2013'), 
('14216','355','6/1/2013'),('14216','739','6/1/2020'),('14216','10','3/1/2021'), 
('14216','1009','6/1/2021')

Answer 1

没有@MatBailie 建议的关于拥有更多结构化数据的建议，这有点令人费解。为完成您的要求，每条记录都需要知道它之前和之后的合同何时生效。我认为您需要尝试排序，因为我不太了解如何对结果进行排序……按 clientid、contractid、日期等？

更新：见cmets。通过 clientid 更改了一些 CTE、JOINS 和 ORDER BY 以更好地进行分区。

CREATE TABLE [client] (
[clientid] [int] NULL,
[contractid] [int] NULL,
[effectivedate] [date] NULL
) ON [PRIMARY]
;

insert into [client] values
('228','2','6/1/2003'),('228','136','6/1/2004'),('228','242','6/1/2008'), 
('228','337','12/1/2012'),('228','584','6/1/2017'),('14216','319','5/1/2013'), 
('14216','355','6/1/2013'),('14216','739','6/1/2020'),('14216','10','3/1/2021'), 
('14216','1009','6/1/2021')
;

--Need a sequence of numbers to create a sequence of fiscal years.
WITH x AS (
  SELECT * FROM (VALUES (0),(1),(2),(3),(4),(5),(6),(7),(8),(9)) as x(a)
), y as (
  SELECT ROW_NUMBER() OVER(ORDER BY tens.a, ones.a) as row_num
  FROM x as ones, x as tens
), fiscYears as (
  SELECT
    fyStart = DATEFROMPARTS(2000 + y.row_num -1, 6, 1)
    , fyEnd = DATEFROMPARTS(2000 + y.row_num, 5, 31)
  FROM y

--Need to order the client records by effective date.
--From updated question... looks like we are reporting by clientid.
), clientOrd as (
  SELECT c2.*, ROW_NUMBER() OVER(PARTITION BY c2.clientid ORDER BY c2.effectivedate) as row_num
      FROM client c2

--For each contract, get the previous and next contracts by effective date.
), clientWNext as (
  SELECT c.*
    , cNext.effectivedate as nextEffectiveDate
    , cPrev.effectivedate as prevEffectiveDate
  FROM clientOrd as c
    LEFT JOIN clientOrd as cNext
      ON cNext.clientid = c.clientid
      AND cNext.row_num = c.row_num + 1
    LEFT JOIN clientOrd as cPrev
      ON cPrev.clientid = c.clientid
      AND cPrev.row_num = c.row_num - 1
)
SELECT
  c.clientid
  , cwn.contractid
  , CASE WHEN cwn.effectiveDate >= fy.fyStart AND cwn.effectiveDate <= fy.fyEnd 
      THEN  cwn.effectivedate
      ELSE null
    END as effectivedate
  , fy.fyStart
FROM fiscYears as fy
  --To get a full FY range for each client, we join to a distinct list of clients.
  JOIN (
    SELECT DISTINCT clientid FROM client
  ) as c
    ON 1=1

  --Need to join the list of contracts.
  INNER JOIN clientWNext as cwn
    ON cwn.clientid = c.clientid
  
    --This is the main join criteria where the effective date is within the fy year start/end.
    AND ((
      cwn.effectivedate >= fy.fyStart
      AND cwn.effectivedate <= fy.fyEnd
    ) 

    --This is the "alternate" join criteria where the previous contrat is still in effect
    --but there is no new contract to supercede the previous.
    OR (
      cwn.prevEffectiveDate < fy.fyStart
      AND cwn.effectiveDate < fy.fyStart
      AND (cwn.nextEffectiveDate > fy.fyEnd OR cwn.nextEffectiveDate IS NULL)
    ))

--Limiting fiscal year date range.
WHERE fy.fyStart >= '1/1/2003'
  AND fy.fyStart < '1/1/2024'
ORDER BY c.clientid, fy.fyStart, cwn.effectivedate

clientid	contractid	effectivedate	fyStart
228	2	2003-06-01	2003-06-01
228	136	2004-06-01	2004-06-01
228	136	null	2005-06-01
228	136	null	2006-06-01
228	136	null	2007-06-01
228	242	2008-06-01	2008-06-01
228	242	null	2009-06-01
228	242	null	2010-06-01
228	242	null	2011-06-01
228	337	2012-12-01	2012-06-01
228	337	null	2013-06-01
228	337	null	2014-06-01
228	337	null	2015-06-01
228	337	null	2016-06-01
228	584	2017-06-01	2017-06-01
228	584	null	2018-06-01
228	584	null	2019-06-01
228	584	null	2020-06-01
228	584	null	2021-06-01
228	584	null	2022-06-01
228	584	null	2023-06-01
14216	319	2013-05-01	2012-06-01
14216	355	2013-06-01	2013-06-01
14216	355	null	2014-06-01
14216	355	null	2015-06-01
14216	355	null	2016-06-01
14216	355	null	2017-06-01
14216	355	null	2018-06-01
14216	355	null	2019-06-01
14216	739	2020-06-01	2020-06-01
14216	10	2021-03-01	2020-06-01
14216	1009	2021-06-01	2021-06-01
14216	1009	null	2022-06-01
14216	1009	null	2023-06-01

fiddle