最大日期/案例陈述

Question

我正在尝试根据其中一列的最大日期拉回一行详细信息。做一个简单的 max(date.field) 不适用于我的表。我最初是通过利用我使用此资源See the entry by Tom H. 构建的此查询来解决该问题的：

SELECT
    HBM_CLIENT.CLIENT_CODE,
    HBM_NAME.NAME AS CLIENT_NAME,
    PART_BILL1.PART_CAT_CODE AS BILLING_CODE,
    PART_BILL1.EMPL_UNO AS BILLING_NAME_UNO,
    HBM_PERSNL_BILL.EMPLOYEE_NAME AS BILLING_NAME,
    PART_BILL1.PERCENTAGE AS BILLING_PERCENTAGE
FROM 
    HBM_CLIENT
INNER JOIN 
    HBM_NAME ON HBM_CLIENT.NAME_UNO = HBM_NAME.NAME_UNO
LEFT OUTER JOIN 
    TBM_CLMAT_PART AS PART_BILL1 ON PART_BILL1.CLIENT_UNO = HBM_CLIENT.CLIENT_UNO
                                 AND PART_BILL1.PART_CAT_CODE = 'BILL'
LEFT OUTER JOIN 
    TBM_CLMAT_PART AS PART_BILL2 ON PART_BILL2.CLIENT_UNO = HBM_CLIENT.CLIENT_UNO
                                 AND PART_BILL2.EFF_DATE > PART_BILL1.EFF_DATE
                                 AND PART_BILL1.PART_CAT_CODE = 'BILL'
LEFT OUTER JOIN 
    HBM_PERSNL AS HBM_PERSNL_BILL ON PART_BILL1.EMPL_UNO = HBM_PERSNL_BILL.EMPL_UNO
GROUP BY 
    HBM_CLIENT.CLIENT_CODE, HBM_NAME.NAME,
    PART_BILL1.PART_CAT_CODE, PART_BILL1.EMPL_UNO,
    HBM_PERSNL_BILL.EMPLOYEE_NAME,
    PART_BILL1.PERCENTAGE,
ORDER BY 
    HBM_CLIENT.CLIENT_CODE

这个查询的问题是，由于某种原因，它并没有拉回所有的结果，原因是因为我使用的解决方案。

我想知道使用 case 语句是否更有意义，但我对 case 语句不是很熟悉。

这是我正在处理的查询：

`SELECT
  HBM_CLIENT.CLIENT_CODE,
  HBM_NAME.NAME AS CLIENT_NAME,
  PART_BILL1.PART_CAT_CODE AS BILLING_CODE,
  PART_BILL1.EMPL_UNO AS BILLING_NAME_UNO,
  HBM_PERSNL_BILL.EMPLOYEE_NAME AS BILLING_NAME,
  PART_BILL1.PERCENTAGE AS BILLING_PERCENTAGE,
  Part_BILL1.EFF_DATE,
 CASE
    WHEN 
    MAX(Part_BILL1.EFF_DATE) > Part_BILL1.EFF_DATE THEN max(Part_BILL1.EFF_DATE) ELSE Part_BILL1.EFF_DATE END  
FROM HBM_CLIENT
INNER JOIN HBM_MATTER
  ON HBM_CLIENT.CLIENT_UNO = HBM_MATTER.CLIENT_UNO
INNER JOIN HBM_NAME
  ON HBM_CLIENT.NAME_UNO = HBM_NAME.NAME_UNO
LEFT OUTER JOIN TBM_CLMAT_PART AS PART_BILL1
  ON PART_BILL1.CLIENT_UNO = HBM_CLIENT.CLIENT_UNO
  AND PART_BILL1.PART_CAT_CODE = 'BILL'
LEFT OUTER JOIN HBM_PERSNL AS HBM_PERSNL_BILL
  ON PART_BILL1.EMPL_UNO = HBM_PERSNL_BILL.EMPL_UNO
WHERE 
(HBM_CLIENT.CLIENT_CODE = '065011') 
GROUP BY
  HBM_CLIENT.CLIENT_CODE,
  HBM_NAME.NAME,
  PART_BILL1.PART_CAT_CODE,
  PART_BILL1.EMPL_UNO,
  HBM_PERSNL_BILL.EMPLOYEE_NAME,
  PART_BILL1.PERCENTAGE,
  Part_BILL1.EFF_DATE
HAVING 
(Part_BILL1.EFF_DATE = CASE
WHEN MAX(Part_BILL1.EFF_DATE) > Part_BILL1.EFF_DATE THEN max(Part_BILL1.EFF_DATE) ELSE Part_BILL1.EFF_DATE END)`

它运行，但它给了我两行而不是一排，即它不只是拉最大。我正在使用 T-SQL 并使用 Visual Studio 编写 SSRS。

我很乐意回答我能回答的任何问题。如果我的解释不是很好，我提前道歉。

| CLIENT_CODE | CLIENT_NAME  | BILLING_NAME | EFF_DATE | PERCENTAGE |
| ------------+--------------+--------------+----------+------------|
| 123456      | Entity, Inc. | Attorney A   | 1/1/1990 |50%         |
| 123456      | Entity, Inc. | Attorney B   | 1/1/1990 |50%         |
| 123456      | Entity, Inc. | Attorney B   | 1/1/2017 |50%         |
| 123456      | Entity, Inc. | Attorney C   | 1/1/2017 |50%         |

Answer 1

这不能通过使用带有ORDER BY 子句的SELECT TOP 1 查询来简化吗？

类似这样的：

SELECT TOP 1
  HBM_CLIENT.CLIENT_CODE,
  HBM_NAME.NAME AS CLIENT_NAME,
  PART_BILL1.PART_CAT_CODE AS BILLING_CODE,
  PART_BILL1.EMPL_UNO AS BILLING_NAME_UNO,
  HBM_PERSNL_BILL.EMPLOYEE_NAME AS BILLING_NAME,
  PART_BILL1.PERCENTAGE AS BILLING_PERCENTAGE,
  Part_BILL1.EFF_DATE,
  ....
FROM HBM_CLIENT
INNER JOIN HBM_MATTER
  ON HBM_CLIENT.CLIENT_UNO = HBM_MATTER.CLIENT_UNO
INNER JOIN HBM_NAME
  ON HBM_CLIENT.NAME_UNO = HBM_NAME.NAME_UNO
LEFT OUTER JOIN TBM_CLMAT_PART AS PART_BILL1
  ON PART_BILL1.CLIENT_UNO = HBM_CLIENT.CLIENT_UNO
  AND PART_BILL1.PART_CAT_CODE = 'BILL'
LEFT OUTER JOIN HBM_PERSNL AS HBM_PERSNL_BILL
  ON PART_BILL1.EMPL_UNO = HBM_PERSNL_BILL.EMPL_UNO
WHERE 
(HBM_CLIENT.CLIENT_CODE = '065011') 
ORDER BY Part_BILL1.EFF_DATE DESC

这应该只为您提供 EFF_DATE 的最新行，而不需要任何特殊的分组或案例逻辑。

编辑：在我看来，获得所需结果的最佳方法是利用 TOP 子句。通常，在编写查询时 像这样，你会写 SELECT TOP 1 或 SELECT TOP 1000，但是 TOPclause 还支持各种功能。您可以使用 子查询、计算、变量以及WITH TIES 获得您想要的功能。

在您的情况下，您希望查询返回任意数量的行，如此长 因为它们都具有相同的EFF_DATE 值，这也将是 表中给定CLIENT_CODE 的最大EFF_DATE 值。

因此，我建议只需将 WITH TIES 参数添加到 TOP 子句以获得您正在寻找的功能：

SELECT TOP 1 WITH TIES
  HBM_CLIENT.CLIENT_CODE,
  HBM_NAME.NAME AS CLIENT_NAME,
  PART_BILL1.PART_CAT_CODE AS BILLING_CODE,
  PART_BILL1.EMPL_UNO AS BILLING_NAME_UNO,
  HBM_PERSNL_BILL.EMPLOYEE_NAME AS BILLING_NAME,
  PART_BILL1.PERCENTAGE AS BILLING_PERCENTAGE,
  Part_BILL1.EFF_DATE,
  ....
FROM HBM_CLIENT
INNER JOIN HBM_MATTER
  ON HBM_CLIENT.CLIENT_UNO = HBM_MATTER.CLIENT_UNO
INNER JOIN HBM_NAME
  ON HBM_CLIENT.NAME_UNO = HBM_NAME.NAME_UNO
LEFT OUTER JOIN TBM_CLMAT_PART AS PART_BILL1
  ON PART_BILL1.CLIENT_UNO = HBM_CLIENT.CLIENT_UNO
  AND PART_BILL1.PART_CAT_CODE = 'BILL'
LEFT OUTER JOIN HBM_PERSNL AS HBM_PERSNL_BILL
  ON PART_BILL1.EMPL_UNO = HBM_PERSNL_BILL.EMPL_UNO
WHERE 
(HBM_CLIENT.CLIENT_CODE = '065011') 
ORDER BY Part_BILL1.EFF_DATE DESC

TOP子句的各种用法的完整解释可以是 找到here.

简单解释一下，WITH TIES 参数意味着您想要 查询以根据值返回数据集的TOP X 行 在ORDER BY 子句中-在您的情况下为EFF_DATE。所以如果有 EFF_DATE 中的两行具有相同值，或者三行，或者五行 一百，它们都会被退回，因为你是有效的 不是搜索TOP 1 ROW，而是搜索 TOP 1 VALUE of EFF_DATE

希望对您有所帮助。