从临时表更新

Question

查询：

SELECT ID, T.c.value('@Address', 'nvarchar(20)' ) as Address
INTO #TMP
FROM TABLE1
    CROSS APPLY XMLData.nodes('/Document') AS T(c)

UPDATE TABLE1
SET HomeAddress = (SELECT TOP 1 t.Address
                   FROM #TMP t
                   WHERE t.ID = ID)

主要是，我需要将数据 OUT 从 XML 字段复制到同一个表中的普通字段。

问题：

1. 为什么所有记录都获得 Table1 上的 HomeAddress？
2. Cursor 真的是更新 Table1 上的值的唯一方法吗？

Answer 1

在更新时，我需要按以下方式完全限定表格：                  

UPDATE TABLE1 SET TABLE1.HomeAddress = (SELECT TOP 1 t.Address  
               FROM #TMP t  
               WHERE t.ID = TABLE1.ID)  

Answer 2

UPDATE T2
SET HomeAddress = t1.address
FROM TABLE2 t2
join TABLE1 t1 on T1.ID = t2.HomeAddressID
and t2.HomeAddress <> t1.address

使用连接。无需临时表或相关子查询。

如果表 1 处于一对多关系中，则这些是处理该问题的一些可能性。如果您有一个值指示一个且只有一个记录（我们的系统中有一个字段选择最重要的地址，它通过触发器维护以保证唯一性），试试这个：

UPDATE T2
SET HomeAddress = t1.address
FROM TABLE2 t2
join TABLE1 t1 on t1.ID = t2.HomeAddressID
WHERE t1.somefield = 'somevalue'
and t2.HomeAddress <> t1.address

如果您需要基于其他字段（例如最近日期）的唯一记录，请尝试以下变体：

UPDATE T2
SET HomeAddress = t1.address
FROM TABLE2 t2
join TABLE1 t1 on t1.ID = t2.HomeAddressID
join (select id, max(somedatefield) from  table1 group by id) t3 on t3.id = t1.id
Where  t2.HomeAddress <> t1.address

Answer 3

SELECT P.TipoComprob,P.NoComprob,C.Importe as ImpIVA1,ROUND(100/P.ImpGravado*C.Importe,1)
 as PorcIVA1 
       INTO #Temporal
       FROM ComprobProv AS P
       LEFT JOIN PasesCompras AS C ON C.TipoComprob=P.TipoAsiento AND 
C.NoComprob=P.NoComprob
WHERE P.PorcIVA1 =0 and P.CatIVA = 'Ri' AND P.ImpGravado>0 AND C.CodCuenta=110303010 AND ROUND(100/P.ImpGravado*C.Importe,1) IN (21.00,10.50,27.00);
go

select * from #Temporal;
go

UPDATE 
   t1 
SET 
   t1.ImpIVA1 =  t2.ImpIVA1, t1.PorcIVA1 = t2.PorcIVA1
FROM
   dbo.ComprobProv t1
   INNER JOIN #Temporal t2
   ON t1.TipoComprob = t2.TipoComprob AND t1.NoComprob = t2.NoComprob;
go

-- 请注意，“GO”对于保留 SQL Server 2017 中的上下文很重要，否则您会发现 “未知字段名称”错误。