Question

我有一张桌子

有5列： -

国家代码
语言
仓库
ActiveFrom
ActiveTo

数据子集（列标题不对齐的道歉： -

Country
Code
   Language 
      Warehouse
                    ActiveFrom          ActiveTo
AT  de  BHU 2011-08-08 00:00:00.000 9999-12-31 23:59:59.000
AT  de  WGN 2011-02-14 00:00:00.000 9999-12-31 23:59:59.000
BE  fr  BHU 2011-09-01 00:00:00.000 9999-12-31 23:59:59.000
BE  fr  WGN 2011-02-14 00:00:00.000 9999-12-31 23:59:59.000
CH  de  WGN 2011-02-14 00:00:00.000 9999-12-31 23:59:59.000
CZ  sk  BHU 2011-08-08 00:00:00.000 9999-12-31 23:59:59.000
CZ  sk  WGN 2011-02-14 00:00:00.000 9999-12-31 23:59:59.000
DE  de  BHU 2011-09-01 00:00:00.000 9999-12-31 23:59:59.000
DE  de  WGN 2011-02-14 00:00:00.000 9999-12-31 23:59:59.000

我想要做的是将ActiveTo列更新为同一仓库的ActiveFrom日期值（给出或花费几毫秒）。

我试过这个： -

update Translations
 set ActiveTo = DateAdd(ms, -3, ot.ActiveFrom)
    from Translations ot
   Where Warehouse = 'WGN'
     and CountryCode = ot.CountryCode

但是，它给出了这些结果。

CountryCode Language    Warehouse   ActiveFrom  ActiveTo
AT  de  BHU 2011-08-08 00:00:00.000 9999-12-31 23:59:59.000
AT  de  WGN 2011-02-14 00:00:00.000 2011-02-13 23:59:59.997
BE  fr  BHU 2011-09-01 00:00:00.000 9999-12-31 23:59:59.000
BE  fr  WGN 2011-02-14 00:00:00.000 2011-02-13 23:59:59.997
CH  de  WGN 2011-02-14 00:00:00.000 2011-02-13 23:59:59.997
CZ  sk  BHU 2011-08-08 00:00:00.000 9999-12-31 23:59:59.000
CZ  sk  WGN 2011-02-14 00:00:00.000 2011-02-13 23:59:59.997
DE  de  BHU 2011-09-01 00:00:00.000 9999-12-31 23:59:59.000
DE  de  WGN 2011-02-14 00:00:00.000 2011-02-13 23:59:59.997

它正在更新正确的行。但价值是错误的，它从自己的ActiveFrom日期而不是其他仓库开始日期中减去3毫秒。

此SQL在派生表中提供正确的结果： -

select t.CountryCode, t.Warehouse, DateAdd (ms, -3, ot.ActiveFrom) as 'TransferToBHU', t.ActiveFrom, t.ActiveTo,
   ot.CountryCode, ot.Warehouse, ot.ActiveFrom, ot.ActiveTo
from Translations t
inner join Translations ot
  on ot.CountryCode= t.CountryCode

t.Warehouse ='WGN' 和ot.Warehouse ='BHU' 按t.CountryCode排序

    CountryCode Warehouse   TransferToBHU   ActiveFrom  ActiveTo    CountryCode Warehouse   ActiveFrom  ActiveTo
AT  WGN 2011-08-07 23:59:59.997 2011-02-14 00:00:00.000 2011-02-13 23:59:59.997 AT  BHU 2011-08-08 00:00:00.000 9999-12-31 23:59:59.000
BE  WGN 2011-08-31 23:59:59.997 2011-02-14 00:00:00.000 2011-02-13 23:59:59.997 BE  BHU 2011-09-01 00:00:00.000 9999-12-31 23:59:59.000
CZ  WGN 2011-08-07 23:59:59.997 2011-02-14 00:00:00.000 2011-02-13 23:59:59.997 CZ  BHU 2011-08-08 00:00:00.000 9999-12-31 23:59:59.000
DE  WGN 2011-08-31 23:59:59.997 2011-02-14 00:00:00.000 2011-02-13 23:59:59.997 DE  BHU 2011-09-01 00:00:00.000 9999-12-31 23:59:59.000

现在，如果我可以将'TransferToBHU'中的计算值转换为更新值，那正是我想要的。

最简单的解决方案是根据'BHU'值更新'WGN'仓库值。但是，最好的更新将基于获取给定CountryCode的两个最新ActiveFrom日期。

任何一种解决方案都是可以接受的，尽管如果有人提供一种适用于日期的解决方案 - 他们会得到接受的答案。

提前致谢。

Answer 1

update t1
set ActiveTo = DateAdd(ms, -3, t2.ActiveFrom)
from Translations t1
   JOIN
   Translations t2 ON t1.CountryCode = t2.CountryCode
Where t1.Warehouse = 'WGN' AND t2.Warehouse = 'BHU'

当然，国家CH没有BHU行，所以：

update t1
set ActiveTo = DateAdd(ms, -3, ISNULL(t2.ActiveFrom,t1.ActiveFrom))
from Translations t1
   LEFT JOIN
   Translations t2 ON t1.CountryCode = t2.CountryCode
Where t1.Warehouse = 'WGN' AND t2.Warehouse = 'BHU'

使用不同行中的值更新SQL表

1 个答案: