find empty holes in attribute (number) ?

Hi,

I have a table with a number and then various attributes.
I'd like to find the 'holes' that is the numbers without any rows
related, at least one of them.

For example if i have :

num   | other attributes
------+-----------------
1     | ....
2     | ....
4     | ....
6     | ....
7     | ....
8     | ....

etc...

I'd like to have a query returning 3 or a list of all holes (3,5).

I was doing :
select min(num +1) FROM table where (num + 1
not in (select num from table));

And it returns me the first free one (3), which is ok for what I do.
Except that today I noticed it's using nearly 1 minute for a table
with 2500 rows to give me a result, and I have to to something
better.

I do have an index on num but explain tells me :
explain select min(num +1) FROM table where
(num + 1 not in (select num from table));
NOTICE:  QUERY PLAN:

Aggregate  (cost=282.06 rows=2426 width=4)
  ->  Seq Scan on table  (cost=282.06 rows=2426 width=4)
          SubPlan
             ->  Seq Scan on table  (cost=282.06 rows=2426
                    width=4)


So the two Seq Scan probably explain the slowness...

Can I improve the SQL query ?
Should I instead process this thing in my program in Perl (since I need only
4 seconds to have the results of select num from table, and i can
easily process that in Perl) ?

If that matters : postgresql 6.5.3 on debian gnu/linux potato.

TIA and regards from France.

--
Patrick.
Because if life has a meaning, we should already know it.