Question

我有这个事实表：

CREATE TABLE [dbo].[FACT_TABLE]
(
    [A_id] [int] NOT NULL,
    [B_id] [int] NOT NULL,
    [C_id] [int] NOT NULL,
    [D_id] [int] NOT NULL,
    [FACT_total] [float] NOT NULL,
    [FACT_average] [float] NOT NULL,

    CONSTRAINT [PK_FACT_TABLE] 
        PRIMARY KEY CLUSTERED ([A_id] ASC, [B_id] ASC,
                               [C_id] ASC, [D_id] ASC)
) ON [PRIMARY]

我有一个T-SQL脚本将数据加载到我的事实表中：

INSERT INTO [dbo].[FACT_TABLE] ([A_id], [B_id], [C_id], [company_id], [D_id],[FACT_total], [FACT_average])
    SELECT 
        b.A_id, c.B_id, d.C_id, e.D_id,
        AVG(FACT_total) AS FACT_total,
        AVG(FACT_average) AS FACT_average
    FROM 
        [dbo].[staging_Area] a
    INNER JOIN 
        [dbo].[DIM_A] b ON a.A_id = b.A_id
    INNER JOIN 
        [dbo].[DIM_B] c ON a.[B_id] = c.B_id
    INNER JOIN 
        [dbo].[DIM_C] d ON a.[C_id] = d.C_id 
    INNER JOIN 
        [dbo].[DIM_D] e ON a.D_id = e.D_id
    GROUP BY  
        b.A_id, c.B_id, d.C_id, e.D_id

如何保证我不会在事实表中插入重复值以避免引用参照完整性错误？

我在该查询之后尝试使用以下语句：

EXCEPT 
SELECT *
FROM [dbo].[FACT_TABLE]

但我认为这不是最佳方法......

任何人都有另一种选择吗？对不起，但我不能把真正的列和表名称......

谢谢！

Answer 1

带有csvListNew.clear()到事实表的

LEFT OUTER JOIN以确保您没有插入已存在的记录：

WHERE FT.Column IS NULL

Answer 2

INSERT INTO [dbo]。[FACT_TABLE]

（[A_id]，[B_id]，[C_id]，[D_id]，[FACT_total]，[FACT_average]）

选择a.A_id，a.B_id，a.C_id，a.D_id，a.FACT_total，a.FACT_average

来自

（选择

b.A_id, c.B_id, d.C_id, e.D_id,
AVG(FACT_total) AS FACT_total,
AVG(FACT_average) AS FACT_average

FROM

[dbo].[staging_Area] a

INNER JOIN

[dbo].[DIM_A] b ON a.A_id = b.A_id

INNER JOIN

[dbo].[DIM_B] c ON a.[B_id] = c.B_id

INNER JOIN [dbo]。[DIM_C] d ON a。[C_id] = d.C_id

INNER JOIN

[dbo].[DIM_D] e ON a.D_id = e.D_id

GROUP BY

b.A_id, c.B_id, d.C_id, e.D_id 

)a

其中

NOT EXISTS

（ SELECT 1 FROM dbao.FACT_TABLE B

WHERE
B.A_id = A.A_id AND B.B_id = A.B_id和b.C_id = a.C_id和b.D_id = a.D_id

事实表中的外键引用完整性 - T-SQL

2 个答案: