Let's say I have a lil ole table like this:

animal amount
cat 12
dog 4
horse 90000


and let's say I want an output table that is the self-join of that
table, but with two things
filtered out:
1 - rows with the same animal (i.e. no rows like:
cat 12 cat 12

2 - either of the rows which contain the same thing but in reverse
order... meaning:
cat 12 dog 4 OR dog 4 cat 12 but not both


Is this possible in a single SQL query or must PL/SQL or program code
be written?

On Feb 4, 8:28 am, "metaperl" <metap... (AT) gmail (DOT) com> wrote:
You can use a theta-join. Example:

select T1.*,T2.* from T T1, T T2 where T1.animal > T2.animal

ANIMAL AMOUNT ANIMAL AMOUNT
---------- ----------- ---------- -----------
dog 4 cat 12
horse 90000 cat 12
horse 90000 dog 4
parrot 4 cat 12
parrot 4 dog 4
parrot 4 horse 90000


/Lennart