Re: problematic query (for me ;-)

"Kim Petersen" <kp(at)kyborg(dot)dk> wrote in message
news:3E660A41(dot)7040509(at)kyborg(dot)dk(dot)(dot)(dot)
> I'm trying to get a hint as to how i can solve this problem:
>
> i have a table:
>
> n | t
> ---+---
> 1 | 2
> 1 | 5
> 2 | 3
> 2 | 5
> 3 | 4
> 3 | 3
> (6 rows)
>
> now i want to find the pairs (n1,n2) where no t's collide - eg in this
> table it would be (1,3) and (3,1).

SELECT DISTINCT nt1.n, nt2.n FROM nt nt1, nt nt2
WHERE NOT EXISTS (
(SELECT t from nt where n=nt1.n) INTERSECT
(SELECT t from nt where n=nt2.n))

n | n
---+---
1 | 3
3 |1
(2 rows)

I'm not sure that it scales well to large datasets...

/Bjarke