使用 JOIN 对 3 个表进行 SQL COUNT

Question

我在一个数据库中有 4 个表。仓库包含客户拥有的盒子，盒子里有文件。有Client 表、Warehouse 表、Boxes 表和Files 表。

所以Client 表有WarehouseID 作为外键，Boxes 表有ClientID 作为外键，Files 表有BoxID 作为外键。我想计算每个客户在我的查询中拥有的盒子和文件的数量，以及进出仓库的盒子数量。 Boxes 和Files 表上的Status 字段确定箱子和文件是进出仓库。我在盒子上运行以下查询并且数字是正确的：

SELECT
    [c].[ClientID],
    [c].[Name] AS [ClientName],
    [w].[Name] AS [WarehouseName],
    COUNT(DISTINCT [b].[BoxID]) AS [BoxCount],
    SUM(CASE WHEN [b].[Status] = @IN THEN 1 ELSE 0 END)) AS [BoxesIn],
    SUM(CASE WHEN [b].[Status] = @OUT THEN 1 ELSE 0 END) AS [BoxesOut],
    SUM(CASE WHEN [b].[DestructionDate] <= GETDATE() THEN 1 ELSE 0 END) AS [BoxesForDestruction],
FROM [Clients] AS [c] INNER JOIN [Boxes] AS [b]
  ON [c].[ClientID] = [b].[ClientID]
INNER JOIN [Warehouses] AS [w]
  ON [c].WarehouseID = [w].[WarehouseID]
WHERE [c].[ClientID] = @ClientID
GROUP BY
  [c].[ClientID],
  [c].[Name],
  [w].[Name]

这会产生以下输出：

ClientID | ClientName | WarehouseName | BoxCount | BoxesIn | BoxesOut | BoxesForDestruction
1        | ACME Corp. | FooFactory    | 22744    | 22699   | 45       | 7888 

计数的输出是正确的。当我将Files 表添加到INNER JOIN 时，数字就会膨胀。这是 SQL：

SELECT
    [c].[ClientID],
    [c].[Name] AS [ClientName],
    [w].[Name] AS [WarehouseName],
    COUNT(DISTINCT [b].[BoxID]) AS [BoxCount],
    COUNT(DISTINCT [f].[FileID]) AS [FileCount], -- *NEW*
    SUM(CASE WHEN [b].[Status] = @IN THEN 1 ELSE 0 END)) AS [BoxesIn],
    SUM(CASE WHEN [b].[Status] = @OUT THEN 1 ELSE 0 END) AS [BoxesOut],
    SUM(CASE WHEN [b].[DestructionDate] <= GETDATE() THEN 1 ELSE 0 END) AS [BoxesForDestruction],
FROM [Clients] AS [c] INNER JOIN [Boxes] AS [b]
  ON [c].[ClientID] = [b].[ClientID]
INNER JOIN [Warehouses] AS [w]
  ON [c].[WarehouseID] = [w].[WarehouseID]
INNER JOIN [Files] AS [f]      -- *NEW*
  ON [b].[BoxID] = [f].[BoxID] -- *NEW*
WHERE [c].[ClientID] = @ClientID
GROUP BY
  [c].[ClientID],
  [c].[Name],
  [w].[Name]

这给了我下面的计数输出（我省略了前 3 列，因为它们不相关）：

BoxCount | FilesCount | BoxesIn | BoxesOut | BoxesForDestruction 
19151    | 411961     | 411381  | 580      | 144615       

FilesCount 是正确的，但其他数字是关闭的。我知道为什么会这样，但我不知道如何解决它。由于在框和文件上的连接返回了多行，因此创建了额外的行。执行SUM 时，额外的行会增加计数。由于仓库只有一行，因此该连接不会影响计数。如何修改我的查询以获取正确数量的进出仓库的文件和盒子？

Answer 1

join 为右侧表格中的每一行重复左侧表格中的每一行。如果您合并多个连接，则某些行将被重复计算。一种解决方案是将计数移动到子查询中。例如：

select  *
from    table1 t1
join    (
        select  table1_id
        ,       count(*)
        from    table2
        group by
                table1_id
        ) t2
on      t2.table1_id = t1.id
join    (
        select  table1_id
        ,       count(*)
        from    table3
        group by
                table1_id
        ) t3
on      t3.table1_id = t1.id

Answer 2

正如 Andomar 所说，除了 Count(*) 之外，我还包括了“as myColumnOne”和“myColumnTwo”，因为它是 SQL Server 2018 所必需的：

 select  *
from    table1 t1
join    (
        select  table1_id
        ,       count(*) as myColumnOne
        from    table2
        group by
                table1_id
        ) t2
on      t2.table1_id = t1.id
join    (
        select  table1_id
        ,       count(*) as myColumnTwo
        from    table3
        group by
                table1_id
        ) t3
on      t3.table1_id = t1.id