如何在使用变量或其他技术的情况下简化子查询 - MS SQL

Question

我仍在学习 SQL，我花了几个小时寻找我的问题的答案，但无济于事。如果这是一个骗局，请指出正确的方向！

我正在根据工作流工具生成的数据构建报告。这是我需要的结果：

| ID | Name | Company | Contract  |
| 1  | John | Acme    | Freelance |
| 2  | Bill | Acme    | Interne   |
| 3  | Mike | Duff    | Salaried  |

这里是主实例表“Instance_Table”：

| ID | Status | Date       |
| 1  | OK     | 2016-10-21 |
| 2  | Failed | 2016-10-24 |
| 3  | OK     | 2016-10-25 |

这是数据所在的“Data_Table”：

| DataID | InstanceID | VariableID | Value      |
| 1000   | 1          | 453        | John       |
| 1001   | 1          | 879        | Acme       |
| 1002   | 1          | 1879       | Freelance  |
| 1003   | 1          | 98         | NULL       |
| 1010   | 2          | 453        | Bill       |
| 1011   | 2          | 879        | Acme       |
| 1012   | 2          | 1879       | NULL       |
| 1013   | 2          | 98         | INT        |
| 1020   | 3          | 453        | Mike       |
| 1021   | 3          | 879        | Duff       |
| 1022   | 3          | 1879       | Salaried   |
| 1023   | 3          | 98         | SAL        |

如您所见，合约有 2 个可能的变量，1879 和 98。根据实例，仅填充一个或两个，并且 98 使用短代码而不是完整的单词。

这是我得到结果的痛苦方式：

SELECT IT.ID,

(SELECT DT.value FROM Data_Table DT
 WHERE DT.VariableID = 453 AND IT.ID = DT.InstanceID) as "Name",

(SELECT DT.value FROM Data_Table DT
 WHERE DT.VariableID = 879 AND IT.ID = DT.InstanceID) as "Company",

CASE
 WHEN (SELECT DT.value FROM Data_Table DT
       WHERE DT.VariableID = 1897 AND IT.ID = DT.InstanceID) is NOT NULL
  THEN (SELECT DT.value FROM Data_Table DT
       WHERE DT.VariableID = 1897 AND IT.ID = DT.InstanceID)
  ELSE CASE
        WHEN (SELECT DT.value FROM Data_Table DT
              WHERE DT.VariableID = 98 AND IT.ID = DT.InstanceID) = 'SAL'
         THEN 'Salaried'
        WHEN (SELECT DT.value FROM Data_Table DT
              WHERE DT.VariableID = 98 AND IT.ID = DT.InstanceID) = 'INT'
         THEN 'Intern'
        WHEN (SELECT DT.value FROM Data_Table DT
              WHERE DT.VariableID = 98 AND IT.ID = DT.InstanceID) = 'FRE'
         THEN 'Freelance'
        ELSE '-'
      END
 END as "Contract"

FROM Instance_Table IT

我在问如何使用变量或其他方法简化每个子查询的重复，因为我可以想象多次查询同一个表会对性能造成巨大影响。

这个例子当然是我实际所做的一个非常简化的版本......我有大约 30 列要放入我的报告中，数据以各种形式出现，有近 70 个子查询实例，其中很多被重复。该查询的运行时间很长，如果能帮助它更快地运行，我们将不胜感激。

Answer 1

大概是这样的吧？

SELECT IT.id,
       DT_Name.value AS NAME,
       DT_Cmpy.value AS Company,
       [Contract]=CASE
                    WHEN DT_Cont_1.value IS NOT NULL THEN DT_Cont_1.value
                    ELSE
                      CASE DT_Cont_2.value
                        WHEN 'SAL' THEN 'Salaried'
                        WHEN 'INT' THEN 'Intern'
                        WHEN 'FRE' THEN 'Freelance'
                        ELSE '-'
                      END
                  END
FROM   instance_table IT
       LEFT JOIN data_table DT_Name
              ON DT_Name.variableid = 453
                 AND IT.id = DT_Name.instanceid
       LEFT JOIN data_table DT_Cmpy
              ON DT_Cmpy.variableid = 879
                 AND IT.id = DT_Cmpy.instanceid
       LEFT JOIN data_table DT_Cont_1
              ON DT_Cont_1.variableid = 1897
                 AND IT.id = DT_Cont_1.instanceid
       LEFT JOIN data_table DT_Cont_2
              ON DT_Cont_2.variableid = 98
                 AND IT.id = DT_Cont_2.instanceid  

Answer 2

使用LEFT JOIN 代替相关子查询

SELECT IT.ID,
       DT1.value                AS "Name",
       DT2.value                AS "Company",
       COALESCE(DT3.value, CASE DT4.value
                             WHEN 'SAL' THEN 'Salaried'
                             WHEN 'INT' THEN 'Intern'
                             WHEN 'FRE' THEN 'Freelance'
                             ELSE '-'
                           END) AS "Contract"
FROM   Instance_Table IT
       LEFT JOIN Data_Table DT1
              ON IT.ID = DT1.InstanceID
                 AND DT1.VariableID = 453
       LEFT JOIN Data_Table DT2
              ON IT.ID = DT2.InstanceID
                 AND DT2.VariableID = 879
       LEFT JOIN Data_Table DT3
              ON IT.ID = DT3.InstanceID
                 AND DT3.VariableID = 1897
       LEFT JOIN Data_Table DT4
              ON IT.ID = DT4.InstanceID
                 AND DT4.VariableID = 98
                 AND DT.value IN ( 'SAL', 'INT', 'FRE' ) 

Answer 3

您可以在不多次连接 data_table 的情况下执行此操作。

SELECT 
    it.ID,
    MAX(CASE WHEN DT.VariableID = 453 THEN DT.VALUE END) AS "Name",
    MAX(CASE WHEN DT.VariableID = 879 THEN DT.VALUE END) AS "Company",
    COALESCE(MAX(CASE WHEN DT.VariableID = 1897 THEN DT.VALUE END), 
        MAX(CASE WHEN DT.VariableID = 98 AND dt.VALUE = 'SAL' THEN 'Salary' 
                 WHEN DT.VariableID = 98 AND dt.VALUE = 'INT' THEN 'Intern'
                 WHEN DT.VariableID = 98 AND dt.VALUE = 'FRE' THEN 'Freelance'
                END)
    ) AS "Contract"
FROM 
    Instance_Table it
    LEFT JOIN Data_Table dt ON  it.ID = dt.InstanceID
GROUP BY it.ID