SQL Server 多对多没有主键关系

Question

我遇到了 T-SQL 问题。该案例有两个表，期望从另一个表中左连接一行并连接不同的行以获得结果。

编辑案例： tabls

我的代码：

SELECT T1.ID,T2.BID,T2.FID,T3.FID,T3.PID,T4.PID,T4.Name,T5.FID
FROM T1
INNER JOIN T2 ON T2.BID = T1.ID
INNER JOIN T3 ON T3.FID = T2.FID
INNER JOIN T4 ON T4.PID = T3.PID
OUTER APPLY(SELECT TOP(1) T5.*,T3.PID FROM T5 INNER JOIN T3 ON T3.FID = T5.FID WHERE T5.BID = T2.BID AND T2.IsDone
 = 1 AND T3.PID = T3.PID ORDER BY NEWID()) T5

我希望 T5.FID 没有重复

我两次得到这个 T5.FID

result

有表格 DDL 和示例插入数据。

CREATE TABLE [dbo].[T1](
    [ID] [int] NOT NULL
) ON [PRIMARY]
GO
CREATE TABLE [dbo].[T2](
    [BID] [int] NOT NULL,
    [FID] [int] NOT NULL,
    [IsDone] [bit] NOT NULL
) ON [PRIMARY]
GO
CREATE TABLE [dbo].[T3](
    [FID] [int] NOT NULL,
    [PID] [int] NOT NULL
) ON [PRIMARY]
GO
CREATE TABLE [dbo].[T4](
    [PID] [int] NOT NULL,
    [Name] [varchar](50) NOT NULL
) ON [PRIMARY]
GO
CREATE TABLE [dbo].[T5](
    [FID] [int] NOT NULL,
    [BID] [int] NOT NULL
) ON [PRIMARY]
GO
INSERT [dbo].[T1] ([ID]) VALUES (1);
INSERT [dbo].[T1] ([ID]) VALUES (2);
INSERT [dbo].[T1] ([ID]) VALUES (3);
INSERT [dbo].[T1] ([ID]) VALUES (4);
INSERT [dbo].[T2] ([BID], [FID], [IsDone]) VALUES (1, 12, 1);
INSERT [dbo].[T2] ([BID], [FID], [IsDone]) VALUES (2, 13, 1);
INSERT [dbo].[T2] ([BID], [FID], [IsDone]) VALUES (2, 14, 1);
INSERT [dbo].[T2] ([BID], [FID], [IsDone]) VALUES (2, 15, 1);
INSERT [dbo].[T2] ([BID], [FID], [IsDone]) VALUES (2, 16, 0);
INSERT [dbo].[T2] ([BID], [FID], [IsDone]) VALUES (3, 17, 0);
INSERT [dbo].[T2] ([BID], [FID], [IsDone]) VALUES (4, 18, 0);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (12, 2);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (13, 3);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (14, 3);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (15, 3);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (16, 3);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (17, 4);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (18, 5);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (19, 3);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (20, 3);
INSERT [dbo].[T3] ([FID], [PID]) VALUES (21, 3);
INSERT [dbo].[T4] ([PID], [Name]) VALUES (2, N'A');
INSERT [dbo].[T4] ([PID], [Name]) VALUES (3, N'B');
INSERT [dbo].[T4] ([PID], [Name]) VALUES (4, N'C');
INSERT [dbo].[T4] ([PID], [Name]) VALUES (5, N'D');
INSERT [dbo].[T5] ([FID], [BID]) VALUES (19, 2);
INSERT [dbo].[T5] ([FID], [BID]) VALUES (20, 2);
INSERT [dbo].[T5] ([FID], [BID]) VALUES (21, 2);

Answer 1

看起来你可以只做DISTINCT，因为Table1 中的数据都完全相同。

SELECT DISTINCT
       T1.AID,
       T1.[Name],
       T2.Property
FROM dbo.Table1 T1
     JOIN dbo.Table2 T2 ON T1.AID = T2.AID;

我个人认为，在一个表中包含多个完全相同的行通常意味着您有需要解决的数据问题。如果确实需要DELETE 重复，我建议看看How to delete duplicate rows in SQL Server?。然后你可以去掉昂贵的DISTINCT 操作符。