SpaceMarine wrote:
> On Nov 9, 4:00 am, havardk <hrkristiansen..._at_hotmail.com> wrote:
>> Ah... Now I understand your problem. Avoiding duplicates complicates
>> matters, because then you will have to traverse the tree the other way
>> to see if any other candidate is listed higher, starting from all rows
>> that satisfy your "start with" criteria.
>>
>> The select below is even slower, but still works for this problem. I
>> suspect there is an easier way to solve this :-)
>
> yeah. i think it may be better just to keep the resultset w/ dupes,
> and then run it thru a scrubber in application code (which is what im
> doing now). my C# method loops thru the resultset (which is usually
> pretty small), and if a row is LEVEL 1, it loops thru the resultset
> again looking for a > LEVEL 1 row. if found, it knows the original-row
> is a child and shouldnt be a root, so it flags it for removal. after
> the initial pass all removals are performed (in a reverse loop to
> preserve row-number integrity), each one cascading down to remove the
> removal-row's children. works.
>
>
> thanks for the suggestions!
>
> sm
Take any query that produces a result set in Oracle.
Wrap it with parentheses so it becomes an in-line view.
SELECT DISTINCT from it.
--
Daniel A. Morgan
Oracle Ace Director & Instructor
University of Washington
damorgan_at_x.washington.edu (replace x with u to respond)
Puget Sound Oracle Users Group
www.psoug.org
Received on Fri Nov 09 2007 - 13:52:36 CST
