;with pivot_t(id, yyyy, val)
as
(
	select 1, 2001, 100 union all
	select 2, 2002, 200 union all
	select 3, 2003, 200 union all
	select 1, 2004, 300 union all
	select 2, 2005, 400 union all
	select 3, 2006, 500 
)
select *
from pivot_t pivot(sum(val) for yyyy in([2001], [2002], [2003])) as p1
	join pivot_t pivot(avg(val) for yyyy in([2004], [2005], [2006])) as p2
		on p1.id = p2.id

/*
id          2001        2002        2003        id          2004        2005        2006
----------- ----------- ----------- ----------- ----------- ----------- ----------- -----------
1           100         NULL        NULL        1           300         NULL        NULL
2           NULL        200         NULL        2           NULL        400         NULL
3           NULL        NULL        200         3           NULL        NULL        500

(3개 행이 영향을 받음)
*/

--drop table #temp
create table #temp(
    그룹 varchar(20)
,   유저수 int
)

insert #temp values 
    ('A', 100)
,   ('B', 200)
,   ('C', 300)
,   ('D', 400)

--pivot/unpivot
;with temp
as
(
    select *
    from #temp
        pivot (sum(유저수) for 그룹 in ([A], [B], [C], [D])) as p1
)
select *
from temp 
    unpivot(유저수 for 그룹 in ([A], [B], [C], [D])) as p1