July 30, 2014 at 3:12 pm
I have an address table, and a log table will only record changes in it. So we wrote a after udpate trigger for it. In our case the trigger only need to record historical changes into the log table. so it only needs to be an after update trigger.
I attached the table create script.
The trigger works fine until a day we found out there are same addresses exist in the log table for the same student.
so below is what I modified the trigger to.
I tested, it seems working OK. Also would like to know do I need to use if not exists statement, or just use in the where not exists like what I did in the following code:
ALTER TRIGGER [dbo].[trg_stuPropertyAddressChangeLog] ON [dbo].[stuPropertyAddress]
FOR UPDATE
AS
DECLARE @rc AS INT ;
SELECT @rc = COUNT(*)
FROM Deleted ;
IF @rc = 0
BEGIN
RETURN ;
END
ELSE
BEGIN
INSERT INTO dbo.stuPropertyAddressChangeLog
( StudentID ,
AddressID ,
ApartmentNum ,
StartDate,
EndDate,
CreateDt ,
CreatedBy
)
SELECT del.StudentID ,
del.AddressID ,
del.ApartmentNum ,
COALESCE(del.ChangeDt,del.CreateDt),
SYSDATETIME(),
SYSDATETIME(),
COALESCE(del.ChangedBy,del.CreatedBy)
FROM deleted del
WHERE NOT EXISTS
(SELECT *
FROM inserted AS ins
JOIN deleted AS del
ON ins.studentid=del.studentid AND ins.AddressID=del.addressid
AND ISNULL(ins.ApartmentNum,'')=ISNULL(del.ApartmentNum,'')
)
END
July 30, 2014 at 6:44 pm
sqlfriends (7/30/2014)
I have an address table, and a log table will only record changes in it. So we wrote a after udpate trigger for it. In our case the trigger only need to record historical changes into the log table. so it only needs to be an after update trigger.I attached the table create script.
The trigger works fine until a day we found out there are same addresses exist in the log table for the same student.
so below is what I modified the trigger to.
I tested, it seems working OK. Also would like to know do I need to use if not exists statement, or just use in the where not exists like what I did in the following code:
ALTER TRIGGER [dbo].[trg_stuPropertyAddressChangeLog] ON [dbo].[stuPropertyAddress]
FOR UPDATE
AS
DECLARE @rc AS INT ;
SELECT @rc = COUNT(*)
FROM Deleted ;
IF @rc = 0
BEGIN
RETURN ;
END
ELSE
BEGIN
INSERT INTO dbo.stuPropertyAddressChangeLog
( StudentID ,
AddressID ,
ApartmentNum ,
StartDate,
EndDate,
CreateDt ,
CreatedBy
)
SELECT del.StudentID ,
del.AddressID ,
del.ApartmentNum ,
COALESCE(del.ChangeDt,del.CreateDt),
SYSDATETIME(),
SYSDATETIME(),
COALESCE(del.ChangedBy,del.CreatedBy)
FROM deleted del
WHERE NOT EXISTS
(SELECT *
FROM inserted AS ins
JOIN deleted AS del
ON ins.studentid=del.studentid AND ins.AddressID=del.addressid
AND ISNULL(ins.ApartmentNum,'')=ISNULL(del.ApartmentNum,'')
)
END
If you don't want dup rows in the log table (dbo.stuPropertyAddressChangeLog), don't you think you should join the deleted table to it not the inserted table in the NOT EXISTS to determine if the data already exists in the log table?
July 31, 2014 at 10:05 am
Thanks, the same address could be in the log table twice in case they move out an address then after a year move back. so this is correct case.
we just don't want same address shows up for the students in two consecutive records.
The reason I join inserted and deleted is to want to make sure is to prevent some one update the table using the same data multiple times. so in this case inserted and deleted are the same.
then I use not exists to exclude for this case.
Does that sound correct?
Thanks
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply