Need query to add duplicates

Question

Guys,
I've been trying to solve the shortest route problem and got stuck in calculating the cumulative distance in a hierarchial query. I've got some idea to deal with it like
sum(distance) over (order by level rows level preceding)But in order for this to happen I may be needing duplicate records. Anyway, let me get straight to what I need.
I have table data like this.
col1      col2       rown
1         10         1
2         20         2
3         30         3
4         40         4
4         70         5
5         80         6
3         90         7
5         95         8I want the result to be like this.
col1      col2
1         10
2         20
3         30
4         40
1         10
2         20
3         30
4         70
5         80
1         10
2         20
3         90
1         10
2         20
3         30
4         70
5         95Please observe that when filling with duplicates, if there are 2 or more occurances of the same, it should be the latest(maximum rown in all). For eg: To fill for 5---95, there are two 3s (3---30 and 3---90) and two 4s(4---40 and 4---70). The latest need to be taken meaning 3---90 and 4---70 are to be taken. I hope I made it clear. Please let me know if you need any further information.
Cheers
Sarma.

Answer 1

Thank you Etbin.
It was line # 16 that was missing. Now it looks okay.
SQL> select finaal.col1, finaal.col2
  2  from (select col1, col2, rown,
  3               row_number() over (partition by col1, rown
  4                                  order by col2 desc) rn
  5        from ( select b.col1 col1,
  6                      b.col2 col2,
  7                      a.rn rown
  8               from sample b, (select col1, col2, rown, rn
  9                               from (select col1, col2, rown, rownum rn,
10                                            lag(col1) over (order by rown) prev
11                                     from sample
12                                    )
13                               where col1 - prev <> 1
14                              ) a
15               where b.col1 < a.col1
16               and b.rown <= a.rown
17               union all
18               select col1,
19                      col2,
20                      rown
21               from sample
22             )
23         ) finaal
24  where finaal.rn = 1
25  order by finaal.rown, finaal.col1
26  /
      COL1       COL2
         1         10
         2         20
         3         30
         4         40
         1         10
         2         20
         3         30
         4         70
         5         80
         1         10
         2         20
         3         90
         1         10
         2         20
         3         90
         4         70
         5         95
         1         15
         1         15
         2         20
         3         90
         4         80
         1         15
         2         20
         3         90
         4         80
         5         95
         6         99
         1         15
         2         20
         3         33
         1         15
         2         20
         3         90
         4         80
         5         97
36 rows selected.
SQL>Cheers
Sarma.