T-SQL 结果集 - 包含 3 个表答案

【问题标题】：T-SQL Result Set - with 3 tablesT-SQL 结果集 - 包含 3 个表
【发布时间】：2019-02-05 03:59:52
【问题描述】：

这是我的代码：

drop table if exists #Table_A
drop table if exists #Table_B
drop table if exists #Table_C

create table #Table_A
(
    [DATE] Date,
    TYPE    INT,
    QTY     INT
);

create table #Table_B
(
    [DATE] Date,
    TYPE    INT,
    QTY     INT
);

create table #Table_C
(
    [DATE] Date,
    TYPE    INT,
    QTY     INT
);

Insert into #Table_A VALUES ('2018-01-01', 1, 10)
Insert into #Table_B VALUES ('2018-01-01', 1, 5)
Insert into #Table_B VALUES ('2018-01-01', 2, 10)
Insert into #Table_B VALUES ('2018-01-01', 3, 10)
Insert into #Table_C VALUES ('2018-01-01', 1, 5)
Insert into #Table_C VALUES ('2018-01-01', 2, 10)
Insert into #Table_C VALUES ('2018-01-01', 3, 10)

查询：

select 
    A.[DATE], 
    SUM(CASE WHEN A.[TYPE] = 1 THEN A.QTY END)  A_QTY,
    SUM(CASE WHEN B.[TYPE] = 1 THEN B.QTY END)  B_QTY,
    SUM(CASE WHEN C.[TYPE] = 1 THEN C.QTY END)  C_QTY,
    SUM(CASE WHEN A.[TYPE] = 2 THEN A.QTY END)  A_QTY,
    SUM(CASE WHEN B.[TYPE] = 2 THEN B.QTY END)  B_QTY,
    SUM(CASE WHEN C.[TYPE] = 2 THEN C.QTY END)  C_QTY,
    SUM(CASE WHEN A.[TYPE] = 3 THEN A.QTY END)  A_QTY,
    SUM(CASE WHEN B.[TYPE] = 3 THEN B.QTY END)  B_QTY,
    SUM(CASE WHEN C.[TYPE] = 3 THEN C.QTY END)  C_QTY
FROM 
    #Table_A A 
    INNER JOIN #Table_B B ON A.DATE = B.DATE AND A.TYPE = B.TYPE
    INNER JOIN #Table_C C ON A.DATE = C.DATE AND A.TYPE = C.TYPE
GROUP BY A.[DATE]

当前结果：

DATE        |   A_QTY   |   B_QTY   |   C_QTY   A_QTY   B_QTY   C_QTY   A_QTY   B_QTY   C_QTY
2018-01-01  |   10      |   5       |   5       NULL    NULL    NULL    NULL    NULL    NULL

预期结果：

DATE        |   A_QTY   |   B_QTY   |   C_QTY   A_QTY   B_QTY   C_QTY   A_QTY   B_QTY   C_QTY    
2018-01-01  |   10      |   5       |   5       NULL    10      10      NULL    10      10

请告诉我如何获得预期的结果。（由于inner join，我知道目前它无法正常工作）。

【问题讨论】：

标签： sql sql-server tsql inner-join

【解决方案1】：

您的查询的问题在于，由于它使用INNER JOINs，因此只计算所有 3 个表中存在的记录/日期元组。

如果有一个表可以始终拥有所有可用的 (TYPE, DATE) 元组（在您的示例数据中，这将是 Table_C），那么您可以使用这是一个起点并使用LEFT JOINs：

FROM #Table_C C 
LEFT JOIN #Table_A A ON A.DATE = C.DATE AND A.TYPE = C.TYPE
LEFT JOIN #Table_B B ON B.DATE = C.DATE AND B.TYPE = C.TYPE

否则，您可能想将您的 INNER JOINs 更改为 FULL OUTER JOINs ：

FROM #Table_A A 
FULL OUTER JOIN #Table_B B ON A.DATE = B.DATE AND A.TYPE = B.TYPE
FULL OUTER JOIN #Table_C C ON A.DATE = C.DATE AND A.TYPE = C.TYPE

FULL OUTER JOIN 解决方案需要调整 SELECT 和 GROUP BY 子句，因为日期可能来自 3 个表中的任何一个。

查询：

SELECT
    COALESCE(A.[DATE], B.[DATE], C.[DATE]) DATE,
    SUM(CASE WHEN A.[TYPE] = 1 THEN A.QTY END)  A_QTY,
    SUM(CASE WHEN B.[TYPE] = 1 THEN B.QTY END)  B_QTY,
    SUM(CASE WHEN C.[TYPE] = 1 THEN C.QTY END)  C_QTY,
    SUM(CASE WHEN A.[TYPE] = 2 THEN A.QTY END)  A_QTY,
    SUM(CASE WHEN B.[TYPE] = 2 THEN B.QTY END)  B_QTY,
    SUM(CASE WHEN C.[TYPE] = 2 THEN C.QTY END)  C_QTY,
    SUM(CASE WHEN A.[TYPE] = 3 THEN A.QTY END)  A_QTY,
    SUM(CASE WHEN B.[TYPE] = 3 THEN B.QTY END)  B_QTY,
    SUM(CASE WHEN C.[TYPE] = 3 THEN C.QTY END)  C_QTY
FROM 
    #Table_A A 
    FULL OUTER JOIN #Table_B B ON A.DATE = B.DATE AND A.TYPE = B.TYPE
    FULL OUTER JOIN #Table_C C ON A.DATE = C.DATE AND A.TYPE = C.TYPE
GROUP BY COALESCE(A.[DATE], B.[DATE], C.[DATE])

这个demo on DB Fiddle返回：

DATE                | A_QTY | B_QTY | C_QTY | A_QTY | B_QTY | C_QTY | A_QTY | B_QTY | C_QTY
:------------------ | ----: | ----: | ----: | ----: | ----: | ----: | ----: | ----: | ----:
01/01/2018 00:00:00 |    10 |     5 |     5 |       |    10 |    10 |       |    10 |    10

【讨论】：