问题
I have this trigger :
CREATE trigger [dbo].[DeriveTheAge] on [dbo].[Student]
after insert,update
as
begin
declare @sid as int;
declare @sdate as date;
select @sid= [Student ID] from inserted;
select @sdate=[Date of Birth] from inserted;
commit TRANSACTION
if(@sdate is not null)
begin
update Student set Age=DATEDIFF(YEAR,@sdate,GETDATE()) where [Student ID]=@sid;
end
print 'Successfully Done'
end
as it suggests, the trigger automatically calculates the Derived attribute "Age" from the date of birth. But I get this error when I do the insert :
(1 row(s) affected)
Successfully Done
Msg 3609, Level 16, State 1, Line 1
The transaction ended in the trigger. The batch has been aborted.
Initially I avoided this error because the rows were getting updated inspite of the error. But now when I am inserting a record from the FORNT END, the record is not updated. Instead, it throws this exception :
Can anyone please help me out?
btw, mine is SQL Server 2008 R2 and Visual Studio 2010.
CORRECTION : The Records are still getting updated. But the Exception is the Vilan.
Update
CREATE TRIGGER [dbo].[DeriveTheAge]
ON [dbo].[Student]
FOR INSERT, UPDATE
AS
BEGIN
UPDATE s
SET Age = DATEDIFF(YEAR, [Date of Birth], CURRENT_TIMESTAMP)
FROM dbo.Student AS s
INNER JOIN inserted AS i
ON s.[Student ID] = i.[Student ID]
WHERE i.[Date of Birth] IS NOT NULL;
commit transaction
END
GO
回答1:
Why are you committing in the trigger? Why are you not handling multi-row inserts or updates? You can't just declare variables and assign them from inserted - what values do you think will get assigned when you update 2, or 15, or 6000 rows?
CREATE TRIGGER [dbo].[DeriveTheAge]
ON [dbo].[Student]
FOR INSERT, UPDATE
AS
BEGIN
UPDATE s
SET Age = DATEDIFF(YEAR, [Date of Birth], CURRENT_TIMESTAMP)
FROM dbo.Student AS s
INNER JOIN inserted AS i
ON s.[Student ID] = i.[Student ID]
WHERE i.[Date of Birth] IS NOT NULL;
END
GO
That all said, why on earth would you need a trigger to calculate someone's age? You can get this from the birth date right now at query time and know that it will be accurate, unlike this stale value you've stored in the table. Note that if their row is not updated for over a year, the age you've put in the table is out of date. When do you go back and update the Age for all rows in the table? Once a day? Anything less and your Age column is completely unreliable and pointless.
Also, DATEDIFF(YEAR
is not a reliable way to calculate age in the first place. All it does is count the number of year boundaries that have been crossed, it has no idea if the person's actual birthday is Jan 1 or Dec 31 or anywhere in between.
Finally, I wouldn't print from the trigger. Who is going to consume that print statement when you're not debugging?
来源:https://stackoverflow.com/questions/7310820/transaction-has-ended-in-trigger-batch-has-been-aborted-derived-attribute