创建缺少键值数据记录的视图

Question

我们有一个包含一些行的表。 我们想创建一个缺少关键列行的视图，并使用最后可用的值填充其他非关键值。

关键列 - 名称、年份和期间。 如果丢失了，我们必须根据以前可用的键集来填充薪水。

表格中的当前数据：

预期数据集：

在这里，以绿色突出显示的行是基于附加的，并且使用最后一个可用的键集填充值。

请求：

知道如何创建视图 我试过了，但没有看到预期的数据。

这是我尝试过的

WITH “EMP”
AS (
  SELECT DISTINCT EMPLOYEE
  FROM EMP_TBL
  )
  ,"QTR"
AS (
  SELECT 'Q1' QUARTER
  
  UNION ALL
  
  SELECT 'Q2' QUARTER
  
  UNION ALL
  
  SELECT 'Q3' QUARTER
  
  UNION ALL
  
  SELECT 'Q4' QUARTER
  )
  ,"YR"
AS (
  SELECT '2015' "YEAR"
  
  UNION ALL
  
  SELECT '2016' "YEAR"
  
  UNION ALL
  
  SELECT '2017' "YEAR"
  
  UNION ALL
  
  SELECT '2018' "YEAR"
  
  UNION ALL
  
  SELECT '2019' "YEAR"
  
  UNION ALL
  
  SELECT '2020' "YEAR"
  
  UNION ALL
  
  SELECT '2021' "YEAR"
  )
  ,”EMP_Y_QTR"
AS (
  SELECT EMPLOYEE
    ,YEAR
    ,QUARTER
  FROM “EMP”
  CROSS JOIN "QTR"
  CROSS JOIN "YR"
  )
  ,"CURR_DATA"
AS (
  SELECT DISTINCT EMPLOYEE
    ,YEAR
    ,QTR
    ,'Y' FLAG
  FROM EMP_FACT_TABLE
  )
  ,"REF_DATA"
AS (
  SELECT A.EMPLOYEE
    ,A.YEAR
    ,A.QUARTER
    ,B.YEAR AS YEAR_TO_BE_APPLIED
    ,B.QTR AS QTR_TO_BE_APPLIED
    ,B.FLAG
  FROM EMP_Y_QTR A
  LEFT JOIN "CURR_DATA" B ON A.EMPLOYEE = B.EMPLOYEE
    AND A.YEAR = B.YEAR
    AND A.QUARTER = B.QTR
  )
  ,REF_DATA_PROC
AS (
  SELECT *
    ,COALESCE(YEAR_TO_BE_APPLIED, (
        LAG(YEAR_TO_BE_APPLIED) OVER (
          ORDER BY EMPLOYEE
            ,YEAR
            ,QUARTER
          )
        ), (
        LEAD(YEAR_TO_BE_APPLIED) OVER (
          ORDER BY EMPLOYEE
            ,YEAR
            ,QUARTER
          )
        )) AS YEAR_APPLY
    ,COALESCE(QTR_TO_BE_APPLIED, (
        LAG(QTR_TO_BE_APPLIED) OVER (
          ORDER BY EMPLOYEE
            ,YEAR
            ,QUARTER
          )
        ), (
        LEAD(QTR_TO_BE_APPLIED) OVER (
          ORDER BY EMPLOYEE
            ,YEAR
            ,QUARTER
          )
        )) AS QUARTER_APPLY
  FROM REF_DATA T
  )
  ,EMP_FACT_TABLE_DATA
AS (
  SELECT DISTINCT [EMPLOYEE]
    ,[CITY]
    ,[YEAR]
    ,[QTR]
    ,[ADDRESS]
    ,[SALARY]
      FROM [EMP_FACT_TABLE]
  )
SELECT [EMPLOYEE]
    ,[CITY]
    ,[YEAR]
    ,[QTR]
    ,[ADDRESS]
    ,[SALARY]
FROM REF_DATA_PROC B
LEFT JOIN EMP_FACT_TABLE_DATA A ON A.EMPLOYEE = B.EMPLOYEE
  AND B.YEAR = A.YEAR
  AND B.QUARTER = A.QTR
WHERE A.EMPLOYEE = B.EMPLOYEE
  AND B.YEAR_APPLY = A.YEAR
  AND B.QUARTER_APPLY = A.QTR
  
  
  
  
  
  
  
  

Answer 1

您可以通过评估一个不同的名称和句点列表来实现，交叉连接它，然后左连接您的初始表 - 但是，根据数据量，这将需要相当长的时间，因为交叉连接通常不是t 选择的工具...无论如何，这里是一个例子：

DECLARE @t TABLE(cName varchar(10), cYear int, cPeriod char(2), cSalary int)

INSERT INTO @t VALUES
('Mark', 2019, 'Q1', 100)
,('Mark', 2019, 'Q2', 200)
,('Mark', 2019, 'Q4', 400)
,('Mark', 2020, 'Q1', 500)
,('Mark', 2020, 'Q3', 600)
,('Mark', 2020, 'Q4', 480)


;WITH ctePeriods AS(
SELECT DISTINCT t.cYear, p.cPeriod
  FROM @t t
  CROSS JOIN (SELECT 'Q1' AS cPeriod UNION ALL SELECT 'Q2' UNION ALL SELECT 'Q3' UNION ALL SELECT 'Q4') p
),
ctePersonPeriods AS(
  SELECT DISTINCT t.cName, p.cYear, p.cPeriod
    FROM @t t
    CROSS JOIN ctePeriods p
),
cteSalary AS(
SELECT p.cName, p.cYear, p.cPeriod, t.cSalary
  FROM ctePersonPeriods p
  LEFT JOIN @t t ON t.cName = p.cName AND t.cYear = p.cYear AND t.cPeriod = p.cPeriod
 )
 SELECT cName, cYear, cPeriod, ISNULL(cSalary, LAG(cSalary) OVER (PARTITION BY cName ORDER BY cYear, cPeriod)) AS cSalary
  FROM cteSalary