Question

Have forgotten SQL queries as have not used it for a long time.

I have a following requirement.
Have a table called match where I keep my competitor details with respect to matches my team have played against them. So some important fields are like this

match_id
competior_id
match_winner_id
ismatchtied
goals_scored_my_team
goals_scored_comp

From this table I want to get the head to head information for all my competitors.
like this

Competitor Matches Wins Losses Draws 
A          10      5    4      1
B          8       3    2      1

Draw information I can get from ismatchtied is set to ‘Y’ or ‘N’.
I want to get all the info from one query. I can get all the info from executing queries separately and do complex logic processing in my server code. But my performance will take a hit.

Any help will be hugely appreciated.

cheers,
Saurav

Answer 1

You could use conditional aggregation, involving CASE expressions inside aggregate functions, like this:

SELECT
  competitor_id,
  COUNT(*) AS Matches,
  COUNT(CASE WHEN goals_scored_my_team > goals_scored_comp THEN 1 END) AS Wins,
  COUNT(CASE WHEN goals_scored_my_team < goals_scored_comp THEN 1 END) AS Losses,
  COUNT(CASE WHEN goals_scored_my_team = goals_scored_comp THEN 1 END) AS Draws
FROM matches
GROUP BY
  competitor_id
;

Every CASE above will evaluate to NULL when the condition isn’t satisfied. And since COUNT(expr) omits NULLs, every COUNT(CASE ...) in the above query will effectively only count rows that match the corresponding WHEN condition.

So, the first COUNT counts only rows where my team scored more against the competitor, i.e. where my team won. In a similar way, the second and the third CASEs get the numbers of losses and draws.