tSQL Trigger Logic, better to have IF clause, or IN clause? - sql

Question

I have been trying to create triggers to lessen the client side code that needs to be written. I have written the following two tSQL triggers and they both seem to produce the same results, I'm just wondering which one is the more proper way to do it. I'm using SQL Server 2012 if that makes any difference.
i.e.
which one uses less resources
executes faster
is more secure against attacks
etc...
CREATE TRIGGER tr_ProductGroup_INSERT_GroupMap
ON [qmgmt].[dbo].[ProductGroup]
After INSERT
AS
BEGIN
if (
select count([inserted].[groupID])
from [inserted]
where [inserted].[groupID] = 1
) = 0
begin
insert into [qmgmt].[dbo].[GroupMap]([parentGroupID], [childGroupID])
select 1, [inserted].[groupID]
from [inserted]
end
END
GO
OR
CREATE TRIGGER tr_ProductGroup_INSERT_GroupMap
ON [qmgmt].[dbo].[ProductGroup]
After INSERT
AS
BEGIN
insert into [qmgmt].[dbo].[GroupMap]([parentGroupID], [childGroupID])
select 1, [inserted].[groupID]
from [inserted]
Where[inserted].[groupID] in
(
select [inserted].[groupID]
from [inserted]
where [inserted].[groupID] <> 1
)
END
GO
UPDATE:
Based on some of the comments here are the inserts I am using. The GroupMap table has the same results no matter which trigger I use.
insert into [qmgmt].[dbo].[ProductGroup]([groupName], [groupDescription]) values ('root', 'The root of all groups')
insert into [qmgmt].[dbo].[ProductGroup]([groupName], [groupDescription]) values ('orphans', 'This is where the members of deleted groups go')
insert into [qmgmt].[dbo].[ProductGroup]([groupName], [groupDescription]) values ('SMGMT', 'Support Management')
insert into [qmgmt].[dbo].[ProductGroup]([groupName], [groupDescription]) values ('ST1', 'Support Tier 1')
insert into [qmgmt].[dbo].[ProductGroup]([groupName], [groupDescription]) values ('ST2', ' Support Tier 2')
insert into [qmgmt].[dbo].[ProductGroup]([groupName], [groupDescription]) values ('ST3', 'Support Tier 3')
insert into [qmgmt].[dbo].[ProductGroup]([groupName], [groupDescription]) values ('SaaSMGMT', 'Express Management')
insert into [qmgmt].[dbo].[ProductGroup]([groupName], [groupDescription]) values ('SaaSSup', 'Support Express')

Answer 1

Since a comment is a bit too small to put this example in, I'm going to put it in an answer. The reason why people say your triggers are functionally different is that, although your test insert row by row, a trigger will also fire when you insert multiple rows into the table in one single operation. Based on your examples you could try the following:
insert into [qmgmt].[dbo].[ProductGroup]([groupName], [groupDescription])
SELECT 'root', 'The root of all groups'
UNION ALL
SELECT 'orphans', 'This is where the members of deleted groups go'
UNION ALL
SELECT 'SMGMT', 'Support Management'
When doing this query the inserted table will hold 3 rows and (depending on the data) the result of the 2 trigger-code-examples might give different results.
Don't worry, this is a common misconception. The rule of thumb with SQL is to always think in record-sets, never in 'a single record with fields'.
As for your question (yes, I'm going for a real answer =)
I would suggest a variation on the second one.
CREATE TRIGGER tr_ProductGroup_INSERT_GroupMap
ON [qmgmt].[dbo].[ProductGroup]
After INSERT
AS
BEGIN
insert into [qmgmt].[dbo].[GroupMap]([parentGroupID], [childGroupID])
select 1, [inserted].[groupID]
from [inserted]
where [inserted].[groupID] <> 1
END
This way the server only needs to run over inserted once, decide which records to 'keep' and then store them right-away into the destination table.
The question now is, if this does what you want it to do...