Question

我使用SQL Anywhere作为我的数据库。

我有两个表SALES DETAIL (POSDETAIL)和库存调整表（AdjustInventory），并创建和查询以显示销售和浪费/调整后的查询。

输出应该是这样的

Product   SalesQty  Value    WastageQty    Value
------------------------------------------------

销售数量将来自POSDETAIL表，WastageQty来自AdjustInventory

POSDETAIL共有435625条记录，AdjustInventory共有183528条记录。

我设计了下面的查询它根据需要给我完美，但问题是查询非常非常慢，在任何日期范围内显示任何特定数据需要大约10到15美元。

此查询是必需的优化，此处调整是查询。

select    
    p.OrderDate, p.ProductId, p.SalesQty,
    f.WastedQty / p.SalesQty as WastedQty,
    p.NetCost, p.EachCost
from
    (select 
         POSDETAIL.PRODNUM as ProductId,
         DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd') as OrderDate, 
         POSDETAIL.NETCOSTEACH as NetCost,POSDETAIL.COSTEACH as EachCost,
         SUM(POSDETAIL.QUAN) as SalesQty
     from 
         DBA.POSDETAIL
     group by 
         POSDETAIL.PRODNUM, POSDETAIL.NETCOSTEACH, POSDETAIL.COSTEACH, 
         DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd')
) as p (ProductId, OrderDate, NetCost, EachCost, SalesQty) 
left outer join
    (select 
         DBA.AdjustInventory.INVENNUM as ProductId,
         DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd') as OrderDate,
         SUM(AdjustInventory.ADJUSTUNITS) as WastedQty
     from 
         DBA.POSDETAIL
     join 
         DBA.AdjustInventory on DBA.POSDETAIL.PRODNUM = DBA.AdjustInventory.INVENNUM
                             and DBA.AdjustInventory.AdjustType = 9 
                             and DATEFORMAT(DBA.AdjustInventory.AdjustTime,'yyyy/mm/dd') = DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd')
     group by 
         DBA.AdjustInventory.INVENNUM, DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd')
    ) as f(ProductId, OrderDate, WastedQty) on p.ProductId = f.ProductId and p.OrderDate = f.OrderDate;

请帮忙

Answer 1

好的，这可能是你的问题。永远不要在嵌套选择中分组。 1.拉出选择，使其成为临时表 2.索引临时表 3.将临时表加入外部查询。

你应该能够在这里做两次，这应该会有所帮助。

Answer 2

not 100% sure of syntax for SQL

select 
         POSDETAIL.PRODNUM as ProductId,
         DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd') as OrderDate, 
         POSDETAIL.NETCOSTEACH as NetCost,POSDETAIL.COSTEACH as EachCost,
         SUM(POSDETAIL.QUAN) as SalesQty
     from 
         DBA.POSDETAIL
     into #P
     group by 
         POSDETAIL.PRODNUM, POSDETAIL.NETCOSTEACH, POSDETAIL.COSTEACH, 
         DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd')

select 
         DBA.AdjustInventory.INVENNUM as ProductId,
         DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd') as OrderDate,
         SUM(AdjustInventory.ADJUSTUNITS) as WastedQty
     from 
         DBA.POSDETAIL
     Into #F
     join 
         DBA.AdjustInventory on DBA.POSDETAIL.PRODNUM = DBA.AdjustInventory.INVENNUM
                             and DBA.AdjustInventory.AdjustType = 9 
                             and DATEFORMAT(DBA.AdjustInventory.AdjustTime,'yyyy/mm/dd') = DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd')
     group by 
         DBA.AdjustInventory.INVENNUM, DATEFORMAT(POSDETAIL.TIMEORD,'yyyy/mm/dd')




select    
    p.OrderDate, p.ProductId, p.SalesQty,
    f.WastedQty / p.SalesQty as WastedQty,
    p.NetCost, p.EachCost
from #p (ProductId, OrderDate, NetCost, EachCost, SalesQty)
Left outer Join #f
     on p.ProductId = f.ProductId and p.OrderDate = f.OrderDate;

自从我使用SQL Anywhere以来已经过了大约16年，很抱歉这是psudeo代码

改进SQL Anywhere查询

2 个答案: