When running this command in SQL Server Management Studio
UPDATE LogChange
SET Created = GETDATE()
WHERE id > 6000
something strange happens.. :)
All rows get the same value in the Created column.
How do I "force" SQL Server to recalculate the GetDate for every row?
A cursor solved it
declare @id int;
Declare toLoop cursor for
select Id from LogChange where id > 6000
OPEN toLoop
FETCH NEXT FROM toLoop INTO @id
WHILE @@FETCH_STATUS = 0
BEGIN
Update LogChange set Created = GETDATE() where id = @id
FETCH NEXT FROM toLoop INTO @id
END
close toLoop
DEALLOCATE toLoop
I have found the following “getdate()” pattern extremely useful over the years. Especially when updating multiple tables in the same transaction. Having the time values exactly match had more benefit to me then being accurate to the millisecond on individual rows.
declare @dtNow datetime
SET @dtNow = GETDATE()
UPDATE LogChange
SET Created = @dtNow
WHERE id > 6000
I don't think there is a way to do them in a single query because the operations in the same clause of a query are evaluated all-at-once.
If you want to guarantee that separate values are applied to each row, write a query that definitely does that:
declare @dt datetime
set @dt = GETDATE()
;With Diffs as (
select
*,
ROW_NUMBER() OVER (ORDER BY id) * 10 as Offset
from
LogChange
WHERE id > 6000
)
UPDATE Diffs
SET Created = DATEADD(millisecond,Offset,@dt)
We know that the above will generate separate offsets for each row, and that we're definitely adding those offsets to a fixed value. It has often been observed that GETDATE() will return the same value for all rows but I cannot find any specific documentation that guarantees this, which is why I've also introduced a @dt variable so that I know it's only assigned once.
