Sql 当左键连接两个表时，Sum似乎是两倍/三倍计数

我试图在一个标识号上连接两个表inventory和sales，并将库存中的货物数量、库存价值和销售价值/数量相加。这是按零件号分组的

以下是我正在做的：

`select inv.[num]
, sum(inv.[Value]) as [Inventory Value]
, sum(inv.[qty]) as [Inventory Qty]
, sum(sales.[Value]) as [Sales Value]
, sum(sales.[qty]) as [Sales qty]
from Inv
left join sales
on inv.[num]=sales.[num]
group by inv.[num]`

库存价值总和大约是50倍，销售额大约是3到4倍。 如何避免过度求和？

使用group by和多对一或多对多的联接时，应首先聚合度量值并在维度上联接

select
    inv_agg.inv_num as 'Inventory Number'
    ,inv_agg.inv_value as 'Inventory Value'
    ,inv_agg.inv_qty as 'Inventory Qty'
    ,sales_agg.sales_value as 'Sales Value'
    ,sales_agg.sales_qty as 'Sales Qty'
from (select 
        inv.num as inv_num
        ,sum(inv.Value) as inv_value
        ,sum(inv.qty) as inv_qty
    from inv
    group by inv.num) inv_agg
left join (select
        sales.num sales_num
        ,sum(sales.Value) as sales_value
        ,sum(sales.qty) as sales_qty
    from sales
    group by sales.num) sales_agg
on inv_agg.inv_num = sales_agg.sales_num

或者，用CTE的

with inv_agg as (select 
        inv.num as inv_num
        ,sum(inv.Value) as inv_value
        ,sum(inv.qty) as inv_qty
    from inv
    group by inv.num),
sales_agg as (select
        sales.num sales_num
        ,sum(sales.Value) as sales_value
        ,sum(sales.qty) as sales_qty
    from sales
    group by sales.num)
select
    inv_agg.inv_num as 'Inventory Number'
    ,inv_agg.inv_value as 'Inventory Value'
    ,inv_agg.inv_qty as 'Inventory Qty'
    ,sales_agg.sales_value as 'Sales Value'
    ,sales_agg.sales_qty as 'Sales Qty'
from inv_agg left join sales_agg
    on inv_agg.inv_num = sales_agg.sales_num

---

正在发生的事情被称为扇风。你能给我们看一些样本数据，这样我们就可以确定哪种方法是最好的修复方法了吗？聚合Inv或单独聚合Sales都可以解决此问题。请尝试添加Sales.num不为null的位置。无论如何，上面的评论给了你正确的答案这是因为inv.[num]=sales.[num]不是一对一的匹配。Inv中的每一行都与sales中的多行配对，您将所有这些组合的[Value]相加。