独特的配对，其中每个值在任一侧仅出现一次

Question

I have this table with two columns p1, p2. I would like a maximum number of combination with the restriction that no value should appear more than once in the output. p1 - p2 is the same as p2 - p1

a - b
a - c
b - c
a - d
e - d
b - d
c - d
e - f
d - f
a - g
e - g
b - g
c - g
d - g
f - g

Expected result:

f - g
d - c
b - a
e nomatchup

This is the data I'm playing with in a simple temp table.

select * 
into #temptable
from (
select 'a' p1,'b' p2 union all
select 'a','c' union all
select 'b','c' union all
select 'a','d' union all
select 'e','d' union all
select 'b','d' union all
select 'c','d' union all
select 'e','f' union all
select 'd','f' union all
select 'a','g' union all
select 'e','g' union all
select 'b','g' union all
select 'c','g' union all
select 'd','g' union all
select 'f','g' 
) z

Think of it like distinct players that need to be allocated to a first round eliminatory tournament and each player gets to play only once in the first round. But because of availability not all combinations are possible. This will be the query for the first day matchups, whoever doesn't get a matchup in the first round will be put in the pool for second day matches. So having a few players with no matchups in the first day is no big deal but I'm trying to maximize the amout of games in day 1 none the less.

I'm thinking I need to do something like concaternating the p1 p2 columns and somehow make it equivalent to p2 p1 and then select distinct from there and separate them again.

Answer 1

You can get matching pairs using a join. This generates all of them:

select concat(t1.p1, ':', t1.p2),
       concat(t2.p1, ':', t2.p2),
       concat(t3.p1, ':', t3.p2)
from t t1 join
     t t2
     on t2.p1 not in (t1.p1, t1.p2) and
        t2.p2 not in (t1.p1, t1.p2) join
     t t3
     on t3.p1 not in (t1.p1, t1.p2, t2.p1, t2.p2) and
        t3.p2 not in (t1.p1, t1.p2, t2.p1, t2.p2);

You can use select top (1) to return an arbitrary single row.

Here is a db<>fiddle.