首先，我介绍了我解决的第一个问题，然后解释了添加到方程式中的其他条件。
我有一个表，该表具有关于居住在房屋中的客户合同的唯一索引，我想知道下一个人要居住一所房屋需要多少天。我们还会在房屋空置时创建“空”合同，因此我们可以通过选择“空”合同来计算错过的美元。
它是这样的： AGREEMENTTYPE = 0表示有人居住 AGREEMENTTYPE = 3表示为空
我现在想知道的是，某人搬出后要重新居住一所房屋（RENTALOBJECTID）需要几天。例如：
RENTALOBJECTID 1, CONTRACTID HC001, AGREEMENTTYPE 0, ended 3/1/2019
RENTALOBJECTID 1, CONTRACTID HC002, AGREEMENTTYPE 3, starts 3/1/2019, ends 6/1/2019 (the empty contract in between).
RENTALOBJECTID 1, CONTRACTID HC003, AGREEMENTTYPE 0, starts 6/1/2019
从HC001到HC003的时间总计为3个月。
我已经有了可以完美地计算出该值的查询，但是现在它开始变得有趣起来。首先，我将向您展示到目前为止的情况：
WITH PMC
AS (SELECT CONTRACTID,
RENTALOBJECTID,
AGREEMENTTYPE,
VALIDFROM,
VALIDTO,
ConcatResult = MIN(CASE
WHEN AGREEMENTTYPE <> 0
THEN FORMAT(VALIDFROM, 'yyyy-MM-dd')
+ FORMAT(ISNULL(VALIDTO, '1900-01-01'), 'yyyy-MM-dd')
+ CONTRACTID
END)
OVER (
PARTITION BY RENTALOBJECTID
ORDER BY VALIDFROM DESC
ROWS BETWEEN UNBOUNDED PRECEDING AND 1 PRECEDING)
FROM PMCCONTRACT)
SELECT *,
NextContractId = SUBSTRING(ConcatResult, 21, 10),
NextValidFrom = CAST(SUBSTRING(ConcatResult, 1, 10) AS DATE),
NextValidTo = CAST(NULLIF(SUBSTRING(ConcatResult, 11, 10), '1900-01-01') AS DATE)
FROM PMC
ORDER BY RENTALOBJECTID,
VALIDFROM
得出以下数据：
CONTRACTID RENTALOBJECTID AGREEMENTTYPE STATECODE VALIDFROM VALIDTO NextContractID NextValidFrom NextValidTo
HC001 1 3 NULL 1/1/2015 3/1/2019 HC004 6/1/2019 NULL
HC002 1 0 2 3/1/2019 4/1/2019 NULL NULL NULL
HC003 1 0 1 4/1/2019 6/1/2019 NULL NULL NULL
HC004 1 3 NULL 6/1/2019 NULL NULL NULL NULL
但是，在某些情况下，AND之间存在多个“ AGREEMENTTYPE = 3”合同，这些合同属于您不希望包含在交货期中的某种类型。我们称之为“ STATECODE”。我们不希望将AGREEMENTTYPE = 3，STATECODE = 2的合同计入天数计算的提前期。
下一个示例显示了我的意思：
RENTALOBJECTID 1, CONTRACTID HC001, AGREEMENTTYPE 0, STATECODE NULL, ended 3/1/2019
RENTALOBJECTID 1, CONTRACTID HC002, AGREEMENTTYPE 3, STATECODE 2, starts 3/1/2019, ends 4/1/2019 (the empty contract in between)
RENTALOBJECTID 1, CONTRACTID HC003, AGREEMENTTYPE 3, STATECODE 1, starts 4/1/2019, ends 6/1/2019
RENTALOBJECTID 1, CONTRACTID HC004, AGREEMENTTYPE 0, STATECODE NULL, starts 6/1/2019
HC001合同与HC004合同之间的时间为3个月，那么HC002合同的-1个月我们应该省略= 2个月的交货时间。
也许您有一个更好的建议，但是现在我希望数据看起来像这样，因为这样我就可以轻松地计算出某个RENTALOBJECTID的合同之间的交货时间（天）。
CONTRACTID RENTALOBJECTID AGREEMENTTYPE STATECODE VALIDFROM VALIDTO NextContractID NextValidFrom NextValidTo NextStatecode2Contract NextStatecode2ValidFrom NextStatecode2ValidTo
HC001 1 3 NULL 1/1/2015 3/1/2019 HC004 6/1/2019 NULL HC002 3/1/2019 4/1/2019
HC002 1 0 2 3/1/2019 4/1/2019 NULL NULL NULL NULL NULL NULL
HC003 1 0 1 4/1/2019 6/1/2019 NULL NULL NULL NULL NULL NULL
HC004 1 3 NULL 6/1/2019 NULL NULL NULL NULL NULL NULL NULL
因此，基本上，我想计算AGREEMENTTYPE 0的CONTRACTS之间的时间，减去AGREEMENTTYPE 3和STATECODE 2的CONTRACTS的时间。但是由于这被用作我的数据模型的维，所以这就是我想要将这最后三列作为计算列。该列的计算应遵循以下方式：
NextStatecode2Contract：
IF CONTRACT has AGREEMENTTYPE 3 and IF one of the next contracts has a CONTRACT with AGREEMENTTYPE 0
AND STATECODE 2 BEFORE the Next Contract with AGREEMENTTYPE 3,
THEN SELECT THE NEXT CONTRACT OF AGREEMENTTYPE 0 AND STATECODE 2.
NextStatecode2ValidFrom：
IF CONTRACT has AGREEMENTTYPE 3 and IF one of the next contracts has a CONTRACT with AGREEMENTTYPE 0
AND STATECODE 2 BEFORE the Next Contract with AGREEMENTTYPE 3,
THEN SELECT VALIDFROM OF THE NEXT CONTRACT OF AGREEMENTTYPE 0 AND STATECODE 2.
NextStatecode2ValidTo
IF CONTRACT has AGREEMENTTYPE 3 and IF one of the next contracts has a CONTRACT with AGREEMENTTYPE 0
AND STATECODE 2 BEFORE the Next Contract with AGREEMENTTYPE 3,
THEN SELECT VALIDTO OF THE NEXT CONTRACT OF AGREEMENTTYPE 0 AND STATECODE 2.