Return 结果基于两个基于 Id 的结果集，不考虑行数

Question

我从两个不同的来源 @Table1, @Table2 接收数据集，table 都有 Id 列。对于相同的 ID，两个来源 return 不同数量的记录。

我想合并基于 @MainTable 的结果，它具有来自 @Table1 和 @Table2 的所有可能 ID。如果另一个 table 没有相同数量的记录，那么它可以与 NULL.

一起应用

示例数据和table设计：

DECLARE @MainTable TABLE (Id INT, [Name] VARCHAR (50));
INSERT INTO @MainTable (Id, [Name]) VALUES (1, 'One'), (2, 'Two'), (3, 'Three');

DECLARE @Table1 TABLE (Id INT, Val1 INT, Val2 INT);
INSERT INTO @Table1 (Id, Val1, Val2) VALUES 
(1, 11, 12), (1, 22, 23), (1, 33, 34), (1, 38, 39), 
(2, 31, 32), (2, 34, 35), 
(3, 44, 45);

DECLARE @Table2 TABLE (Id INT, Val3 INT, Val4 INT);
INSERT INTO @Table2 (Id, Val3, Val4) VALUES 
(1, 20, 21), (1, 25, 26), 
(2, 30, 31), (2, 33, 34), (2, 36, 37), (2, 38, 39), (2, 40, 41),
(3, 51, 52), (3, 54, 55), (3, 56, 57), (3, 58, 59);

预期输出：

Id	Name	Val1	Val2	Val3	Val4
1	One	11	12	20	21
1	One	22	23	25	26
1	One	33	34	NULL	NULL
1	One	38	39	NULL	NULL
2	Two	31	32	30	31
2	Two	34	35	33	34
2	Two	NULL	NULL	36	37
2	Two	NULL	NULL	38	39
2	Two	NULL	NULL	40	41
3	Three	44	55	51	52
3	Three	NULL	NULL	54	55
3	Three	NULL	NULL	56	57
3	Three	NULL	NULL	58	59

我正在使用 SQL Server 2012。我已经尝试了下面的一些代码，但没有帮助：

SELECT MT.Id, MT.[Name], A.Val1, A.Val2, B.Val3, B.Val4
FROM @MainTable MT
OUTER APPLY ( SELECT * FROM @Table1 T1 WHERE T1.Id = MT.Id ) A
OUTER APPLY ( SELECT * FROM @Table2 T2 WHERE T2.Id = MT.Id ) B

SELECT *
FROM @MainTable MT
LEFT OUTER JOIN @Table1 T1 ON T1.Id = MT.Id
LEFT OUTER JOIN @Table2 T2 ON T2.Id = MT.Id

Answer 1

在行之间读取，但如果您不关心哪些行连接到哪些行（因为我看不到 @Table1 到 @Table2 中的行之间没有关系），您可以这样做:

WITH T1 AS(
    SELECT T1.Id,
           T1.Val1,
           T1.Val2,
           ROW_NUMBER() OVER (PARTITION BY T1.Id ORDER BY (SELECT NULL)) AS RN --Define your ORDER BY properly here
    FROM @Table1 T1),
T2 AS(
    SELECT T2.Id,
           T2.Val3,
           T2.Val4,
           ROW_NUMBER() OVER (PARTITION BY T2.Id ORDER BY (SELECT NULL)) AS RN --Define your ORDER BY properly here
    FROM @Table2 T2),
RNs AS(
    SELECT T1.Id,
           T1.RN
    FROM T1
    UNION
    SELECT T2.Id,
           T2.RN
    FROM T2)
SELECT MT.Id,
       T1.Val1,
       T1.Val2,
       T2.Val3,
       T2.Val4
FROM @MainTable MT
     JOIN RNs ON MT.Id = RNs.Id
     LEFT JOIN T1 ON RNs.Id = T1.ID
                 AND RNs.RN = T1.RN
     LEFT JOIN T2 ON RNs.Id = T2.ID
                 AND RNs.RN = T2.RN;

db<>fiddle

Answer 2

Larnu 先于我。但我仍在发布我的答案，因为我认为如果数据集变大，我的答案会更快一些。

DECLARE @MainTable TABLE (Id INT, [Name] VARCHAR (50));
INSERT INTO @MainTable (Id, [Name]) VALUES (1, 'One'), (2, 'Two'), (3, 'Three');

DECLARE @Table1 TABLE (Id INT, Val1 INT, Val2 INT);
INSERT INTO @Table1 (Id, Val1, Val2) VALUES 
(1, 11, 12), (1, 22, 23), (1, 33, 34), (1, 38, 39), 
(2, 31, 32), (2, 34, 35), 
(3, 44, 45);

DECLARE @Table2 TABLE (Id INT, Val3 INT, Val4 INT);
INSERT INTO @Table2 (Id, Val3, Val4) VALUES 
(1, 20, 21), (1, 25, 26), 
(2, 30, 31), (2, 33, 34), (2, 36, 37), (2, 38, 39), (2, 40, 41),
(3, 51, 52), (3, 54, 55), (3, 56, 57), (3, 58, 59);


with cte1 as (
    select
        t1.*,
        row_number() over (partition by t1.Id order by t1.Val1) as seqno 
    from @Table1 t1
),
cte2 as (
    select
        t2.*,
        row_number() over (partition by t2.Id order by t2.Val3) as seqno 
    from @Table2 t2
),
cte3 as (
    select isnull(c1.Id, c2.Id) as Id, c1.Val1, c1.Val2, c2.Val3, c2.Val4
    from cte1 c1
        full outer join cte2 c2 on (c2.Id = c1.Id and c2.seqno = c1.seqno)
)
select
    m.Id,
    m.[Name],
    c3.Val1,
    c3.Val2,
    c3.Val3,
    c3.Val4
from @MainTable m
    left outer join cte3 c3 on (c3.Id = m.Id);

Return 结果基于两个基于 Id 的结果集，不考虑行数

Return result based on two results set based on Id, irrespective of rows count

sql

sql-server

sql-server-2012