Insert Update Duplicate Value SQL - sql

Question

I have this stored procedure:
alter procedure spGroupInsert
(#Group varchar(5))
as
if not exists (select * from tbGroup where Group = #Group)
begin
insert into tbGroup(Group)
values(#Group)
end
else
begin
waitfor delay '00:00:01'
end
The stored procedure is designed to prevent duplicates on tbGroup. Next, I need to have an UPDATE stored procedure:
alter procedure spGroupUpdate
(#GroupID int, #Group varchar(5))
as
begin
update tbGroup
set Group = #Group
where GroupID = #GroupID
end
The table should be:
GroupID Group
1 A
2 B
3 C
4 D
ff.
GroupID is identity. For Insert SP, I am really sure there won't be a problem.
But, if I execute the Update stored procedure, then I change the Group. It will be a duplicate. For instance, if I update Group A to B. Then it will make a duplicate of B.
How I can prevent this in T-SQL in my Update stored procedure?
Thank you.

Answer 1

ALTER TABLE tbGroup
ADD CONSTRAINT UC_Group UNIQUE (Group);
this will handle that in the update stored procedure. It will not update when there is an existing row with the same Group
alter procedure spGroupUpdate
(#GroupID int, #Group varchar(5))
as
begin
update tbGroup
set Group = #Group
WHERE GroupID = #GroupID
AND NOT EXISTS
(
SELECT *
FROM tbGroup x
WHERE x.GroupID <> #GroupID
AND x.Group = #Group
)
end

Answer 2

You are saying that for spGroupInsert you are sure that there won't be a problem. In fact, there is a problem.
This procedure doesn't guarantee that you'll never insert a duplicate. If two sessions are trying to insert the same value at the same time, you can easily get duplicates.
Both sessions can do the check if not exists at the same time and both can proceed with INSERT.
alter procedure spGroupInsert
(#Group varchar(5))
as
if not exists (select * from tbGroup where Group = #Group)
begin
insert into tbGroup(Group)values(#Group)
end
else
begin
waitfor delay '00:00:01'
end
The only way to guarantee that Group values are unique is to create a unique constraint, which is usually implemented as a unique index.
CREATE UNIQUE NONCLUSTERED INDEX [IX_Group] ON [dbo].[tbGroup]
(
[Group] ASC
)
GO
With such unique index in place one of the sessions from the example above would fail to INSERT the duplicate and the caller of the stored procedure would receive an error message about unique constraint violation. The caller would need to decide what to do about this error, how to handle it.
The check if not exists reduces the chances of getting this error, but it can't prevent it completely. So, with the check if not exists or without the check if not exists the INSERT can fail and your code should be able to handle this situation.
It seems that you are happy to suppress/ignore the error. In this case a simple TRY ... CATCH can be enough.
alter procedure spGroupInsert
(#Group varchar(5))
as
BEGIN
SET NOCOUNT ON; SET XACT_ABORT ON;
BEGIN TRANSACTION;
BEGIN TRY
if not exists (select * from tbGroup where Group = #Group)
begin
insert into tbGroup(Group)values(#Group);
end
COMMIT TRANSACTION;
END TRY
BEGIN CATCH
ROLLBACK TRANSACTION;
END CATCH;
END
With this stored procedure (and unique index), if two sessions try to call it with the same Group value at the same time, only one will actually insert the value and the second would silently fail and do nothing. The caller would not know about this collision. In your case it may be acceptable.