Question

I want to simply overwrite values in a column of a table with values from a column in another table. I have a table based off from another table without a unique identifier in one of the columns, so I don't want to use joins, but just update the values since the rows are in the same order. How do I do that? So far I have tried two different approaches where Approach A only put the value from the first row into every row of the updated table whereas Approach B does not work at all.

Approach A:
Update Transactions
SET Transactions.Amount = Transactions_raw.Amount
FROM Transactions_raw


Approach B:
UPDATE Transactions
SET Amount = (SELECT Amount FROM Transactions_raw)

Answer 1

您需要某种join来匹配表-即使使用人工键也是如此：

update t
    set Amount = tr.Amount
from (select t.*, row_number() over (order by (select null)) as seqnum
      from Transactions t
     ) t join
     (select tr.*, row_number() over (order by (select null)) as seqnum
      from Transactions_raw tr
     ) tr
     on t.seqnum  = tr.seqnum;

Answer 2

您假设行的顺序相同可能会误导您。如果您在没有select的情况下执行order by语句，并且在两个表中都看到相同的顺序，那么这不是您要依靠的。不能保证所谓的订单。相反，您必须具有一些订购规则。有了此规则后，您可以将其放在order by中，然后可以根据此顺序将ID列添加到两个表中。

您可以使用以下语句计算ID值：

update Transactions 
set Id = row_number() over(order by ...)

然后，您可以使用常规的inner join。

How can I overwrite a column from a column in another table without a join?

2 个答案: