Question

我有2张桌子。一个是销售表（header），另一个是销售线（lines）。

我想要做的是找出销售表或销售线中的任何记录是否被修改。如果它们被修改，我想将整个销售订单插入另一个表中。我在数据库ModifiedDateTime中使用了一列（销售行中的更改不会影响salestable中的修改日期，反之亦然）。

例如 - 销售表：

SalesID DeliveryAddress ModifiedDatetime
   1        Street1       2017-04-20 2:00:00
   2        Street3       2017-04-19 3:00:00

销售热线

SalesID ItemId Qty ModifiedDateTime
 1       100    2     2017-04-20 3:00:00
 1       101    3     2017-04-20 2:00:00
 2       100    5     2017-04-19 1:00:00

所以我有一个存储过程，它从这些表中选择记录并将它们插入到其他表中。它使用修改的日期时间来了解哪些记录已被修改。我想要做的是将整个订单及其行插入到集成表中。

我现在所做的是

Insert into IntegrationSalesTable (salesid, deliveryaddress, modifieddatetime)
    select salesid, deliveryaddress, modifieddatetime 
    from salestable 
    where modifieddatetime > 2017-04-20 2:30:00

对于销售线

 Insert into IntegrationSalesLine(salesid, itemid, qty, modifieddatetime)
     select salesid, itemid, qty, modifieddatetime 
     from salestable 
     where modifieddatetime > 2017-04-20 2:30:00

现在，我得IntegrationSalesTable为空 - 而IntegrationSalesLine有这一行：

SalesID ItemId Qty ModifiedDateTime
  1       100    2     2017-04-20 3:00:00

我想要实现的目标：

IntegrationSalesTable：

  SalesID DeliveryAddress ModifiedDatetime
   1        Street1       2017-04-20 2:00:00

IntegrationSalesLine：

SalesID ItemId Qty ModifiedDateTime
 1       100    2     2017-04-20 3:00:00
 1       101    3     2017-04-20 2:00:00

我知道我需要在行和表中选择所有已修改的销售订单，然后使用该数据运行查询。但我无法想办法。有什么建议我怎样才能达到这个结果？

Answer 1

这应该做你需要的。它会选择SalesID的所有条目，salesline中的日期与salestable中的日期不匹配

use testdb;

-- This stores all the SalesID's where something changed
DECLARE @modsalesids TABLE
(
   SalesID INT 
)

-- Select SalesID's where dates are different
insert into @modsalesids(SalesID)
select st.SalesID from salestable st
join salesline sl on st.SalesID=sl.SalesID 
where 
-- use '<>' if you want to include _all_ changed dates, or > when you want only changes where the item list changed after the sale
    sl.ModifiedDateTime <> st.ModifiedDateTime 


-- Remove duplicate SalesID's, make sure to keep the ';' 
;

WITH CTE AS
(
SELECT *,ROW_NUMBER() OVER (PARTITION BY SalesID ORDER BY SalesID) AS RN
FROM @modsalesids
)

DELETE FROM CTE WHERE RN<>1

-- Select records for the modified salesIds
insert into integrationsalesline(SalesID, ItemId, Qty, ModifiedDateTime) 
select SalesID, ItemId, Qty, ModifiedDateTime from salesline where SalesID in (SELECT SalesID from @modsalesids);
insert into integrationsalestable(SalesID, DeliveryAddress, ModifiedDateTime) 
select SalesID, DeliveryAddress, ModifiedDateTime from salestable where SalesID in (SELECT SalesID from @modsalesids);

-- Debug stuff

select * from @modsalesids;

select * from integrationsalesline;
select * from integrationsalestable;

如果您感到好奇，这里是删除重复项的部分的source。这就是我设置测试用例的方法：

-- create testdb;

--create table salestable (
--  SalesID INT,
--  DeliveryAddress VARCHAR(200),
--  ModifiedDateTime VARCHAR(200)
--)

--create table salesline (
--   SalesID INT,
--   ItemId INT,
--   Qty INT,
--   ModifiedDateTime VARCHAR(200)
--)

--create table integrationsalestable (
--  SalesID INT,
--  DeliveryAddress VARCHAR(200),
--  ModifiedDateTime VARCHAR(200)
--)

--create table integrationsalesline (
--   SalesID INT,
--   ItemId INT,
--   Qty INT,
--   ModifiedDateTime VARCHAR(200)
--)

--insert INTO salestable  values(1,'Street1','2017-04-20 2:00:00')
--insert INTO salestable  values(2,'Street3','2017-04-20 3:00:00')

--insert INTO salesline values(1, 100, 2, '2017-04-20 3:00:00')
--insert INTO salesline values(1, 107, 7, '2017-04-20 3:00:00')
--insert INTO salesline values(1, 101, 3, '2017-04-20 2:00:00')
--insert INTO salesline values(2, 100, 5, '2017-04-20 1:00:00')

查询结果SQL

1 个答案: