Question

I have a table with the following data

key     code
============
key1    220
key1    221
key2    220
key2    221
key2    222
key3    220
key4    255
key5    220
key5    221

A combination of key and code cannot repeat in the table (the 2 columns are used as composite primary keys).
I will like a query that will select the matching keys for a given set of codes e.g. for codes 220, 221, the expected result will be key1 and key 5 because only these match exactly these code set.

Answer 1

Insert the values you are looking for to a table variable @Match and use that in a query against your table @T.

declare @T table
(
  [key] char(4),
  code int,
  primary key([key], code)
)

insert into @T values
('key1',    220),
('key1',    221),
('key2',    220),
('key2',    221),
('key2',    222),
('key3',    220),
('key3',    333),
('key4',    255),
('key5',    220),
('key5',    221)

declare @Match table (code int)
insert into @Match values (220),(221)

select T.[key]
from @T as T
  left outer join @Match as M
    on T.code = M.code
group by T.[key]
having count(*) = (select count(*) from @Match) and
       count(M.code) = (select count(*) from @Match)

https://data.stackexchange.com/stackoverflow/q/119432/