比较Oracle 11g中2个不同行的2个不同列

Question

I have computed a table from a query and it looks like this:

UID1 UID2
2    3
2    15
3    2
7    12
12   7
15   2

I only need the unique tuples here. ie Out of the tuples where UID1=2,UID2=3 and UID1=3,UID2=2, I need only 1 tuple in the output.
Tried using join on this table with UID values swapped in the other table, but again the same result appears in the output.
Any suggestions, please?

Answer 1

See this SQL Fiddle that should solve your problem.

It's a self-join on both combinations of uid1 = uid2 and keeps only 1 result based on whichever uid1 is "smaller".

select 
  t.*
from
  test t
  inner join test t2
    on t.uid1 = t2.uid2 
       and t.uid2 = t2.uid1
where
  t.uid1 < t2.uid1;