IF EXISTS
(
   SELECT 1
   FROM dbo.CompanyAddress
   WHERE CompanyID = @CompanyID
   AND Address = @AddressID
)
BEGIN
   RETURN;
END
INSERT dbo.CompanyAddress
(
   CompanyID,
   AddressID
)
SELECT
   @CompanyID,
   @AddressID;

On 7/17/08 4:18 PM, in article
3D798A8D-18D5-41B0-83DD-F5539195FDD0@microsoft.com, "Sam"

if not exists ( select null from CompanyAddresses where CompanyID =
@CompanyID and  AddressID = @AddressID)
begin

INSERT INTO CompanyAddresses
 (CompanyID, AddressID)
 VALUES (@CompanyID, @AddressID)
end

Also consider IGNORE_DUP_KEY option, which allows you to not insert
duplicates without raising an error.

When I'm writing data deployment scripts, I often need to perform both
inserts or updates, depending on if the row already happens to exists on the
target server at the time the script is run in production. I hate duplicating
stuff, so I first insert the data into a table variable. Once done, I can
perform updates and inserts on the target table by joining to the table
variable. It works like a charm and really simplifies the script. Pardon the
pseudo code, but y'all get the idea.

-- First insert data into a table variable:
declare @I table ( a,b,c )

insert into @I ( a,b,c ) select a,b,c
insert into @I ( a,b,c. ) select a,b,c
insert into @I ( a,b,c. ) select a,b,c

-- Next, update rows that do exist:
update T
set T.a = @I.a, ...
from T
join @I on T.pk = @I.pk

-- Finally, insert rows that don't exist:
insert into T ( a,b,c )
select a,b,c
from @I
left join T on T.pk = @I.pk
where T.pk is null