Calculate won, tie and lost games in postgresql - postgresql

Question

I have two tables "matches" and "opponents".
Matches
id | date
---+------------
1 | 2016-03-21 21:00:00
2 | 2016-03-22 09:00:00
...
Opponents
(score is null if not played)
id | match_id | team_id | score
---+----------+---------+------------
1 | 1 | 1 | 0
2 | 1 | 2 | 1
3 | 2 | 3 | 1
4 | 2 | 4 | 1
4 | 3 | 1 |
4 | 3 | 2 |
....
The goal is to create the following table
Team | won | tie | lost | total
-----+-----+-----+------+----------
2 | 1 | 0 | 0 | 1
3 | 0 | 1 | 0 | 1
4 | 0 | 1 | 0 | 1
1 | 0 | 0 | 1 | 1
Postgres v9.5
How do I do this? (Im open to maybe moving the "score" to somewhere else in my model if it makes sense.)

Answer 1

Divide et impera my son
with teams as (
select distinct team_id from opponents
),
teamgames as (
select t.team_id, o.match_id, o.score as team_score, oo.score as opponent_score
from teams t
join opponents o on t.team_id = o.team_id
join opponents oo on (oo.match_id = o.match_id and oo.id != o.id)
),
rankgames as (
select
team_id,
case
when team_score > opponent_score then 1
else 0
end as win,
case
when team_score = opponent_score then 1
else 0
end as tie,
case
when team_score < opponent_score then 1
else 0
end as loss
from teamgames
),
rank as (
select
team_id, sum(win) as win, sum(tie) as tie, sum(loss) as loss,
sum( win * 3 + tie * 1 ) as score
from rankgames
group by team_id
order by score desc
)
select * from rank;
Note1: You probably don't need the first "with" as you probably have already a table with one record per team
Note2: i think you can also achieve the same result with one single query, but in this way steps are clearer