with t(ikey, position, change_date) as
(
  select 1, 1, '04.05.2017 10:59:09' from dual
  union all
  select 2, 1, '05.05.2017 10:59:09' from dual
  union all
  select 3, 1, '06.05.2017 10:59:09' from dual
  union all
  select 4, 1, '08.05.2017 10:59:09' from dual
  union all
  select 5, 2, '04.05.2017 10:59:09' from dual
  union all
  select 6, 3, '05.05.2017 10:59:09' from dual
  union all
  select 7, 3, '06.05.2017 10:59:09' from dual
  union all
  select 8, 3, '09.05.2017 10:59:09' from dual
  union all
  select 9, 3, '07.05.2017 10:59:09' from dual
  union all
  select 10, 1, '07.05.2017 10:59:09' from dual
)
select ikey,
  position,
  change_date,
  lead(change_date, 1) over(partition by position order by change_date) as next_change_date
from t
order by ikey

with t(customer, tran_date) as
(
  select 1, to_date('01.01.2013', 'dd.mm.yyyy') from dual
  union all
  select 2, to_date('05.01.2013', 'dd.mm.yyyy') from dual
  union all
  select 2, to_date('06.01.2013', 'dd.mm.yyyy') from dual
  union all
  select 2, to_date('07.01.2013', 'dd.mm.yyyy') from dual
  union all
  select 2, to_date('04.03.2013', 'dd.mm.yyyy') from dual
  union all
  select 2, to_date('21.02.2013', 'dd.mm.yyyy') from dual
  union all
  select 3, to_date('15.03.2013', 'dd.mm.yyyy') from dual
  union all
  select 3, to_date('16.01.2013', 'dd.mm.yyyy') from dual
  union all
  select 3, to_date('17.02.2013', 'dd.mm.yyyy') from dual
)
select *
from
(
  select customer,
    tran_date,
    row_number() over(partition by customer order by tran_date desc) as rn
  from t
)
where rn in (1, 2)
order by customer, tran_date