declare @st table (id int, id2 int, dte datetime)
insert into @st
select 1, 102, '03/03/2006' union all
select 2, 109, '10/03/2006' union all
select 3, 122, '05/01/2005' union all
select 4, 134, '02/23/2005'

declare @dy table (id2 int, de_id int, dte2 datetime)
insert into @dy
select 102, 33, '04/10/2007' union all
select 102, 34, '03/01/2006' union all
select 102, 34, '03/01/2006' union all
select 102, 35, '02/10/2002' union all
select 102, 36, '11/28/1998' union all
select 109, 39, '01/12/2007' union all
select 109, 40, '02/03/2007' union all
select 122, 30, '05/18/1998' union all
select 122, 32, '02/03/1994' union all
select 122, 33, '11/27/1992' union all
select 134, 15, '01/08/2005' union all
select 134, 16, '09/05/2005'

id   de_id  dte2
-----------------------
1    34    '03/01/2006' <- this is the highest dte2 for 109, lower than dte for it
2    NULL  NULL         <- no record in @dy with dte2 lower than dte
3    30    '05/18/1998'
4    15    '01/08/2005'

select st.id, dy.dte2
from @st as st
left join (
	select x.id, max(x.dte2) as dte2 from (
		select distinct st1.id, d1.dte2
		from @st as st1
		left join @dy as d1
		on st1.id2=d1.id2
		where d1.dte2 < st1.dte
	) as x
	group by x.id
) as dy
on st.id = dy.id

select b.id, c.de_id, b.dte2
from (
	select st.id, st.id2, dyX.dte2
	from @st as st
	left join (
		select x.id, max(x.dte2) as dte2 from (
			select distinct st1.id, d1.dte2
			from @st as st1
			left join @dy as d1
			on st1.id2=d1.id2
			where d1.dte2 < st1.dte
		) as x
		group by x.id
	) as dyX
	on dyX.id=st.id
) as b
left join (
	select distinct d.id2, d.dte2, e.de_id
	from (
		select id2, dte2
		from @dy
	) as d
	left join @dy as e
	on d.id2=e.id2 and d.dte2=e.dte2
) as c
on c.id2=b.id2 and c.dte2=b.dte2