Problem Statement:

I am working on this simple dataset from Kaggle. I have provided a snippet of data with only required columns in below table. Dataset is quite simple, it has all IPL (cricket) matches listed with teams who played each match (team1 and team2) along with winner of that match.

Now I am trying to get total matches played by all teams along with matches won by each team, I have again provided a snippet of output below the code. Same can be performed by "finding all occurrences of a particular team in column team1" + "finding all occurrences of a particular team in column team2".

While the code does give proper result, I can sense this is not the best approach. I would like to know some better way to do it along with good practices and naming conventions to follow.


team1 team2 winner
Royal Challengers Bangalore Kolkata Knight Riders Kolkata Knight Riders
Kings XI Punjab Chennai Super Kings Chennai Super Kings
Delhi Daredevils Rajasthan Royals Delhi Daredevils
Mumbai Indians Royal Challengers Bangalore Royal Challengers Bangalore
Kolkata Knight Riders Deccan Chargers Kolkata Knight Riders
Rajasthan Royals Kings XI Punjab Rajasthan Royals


SELECT t1.team1 AS team, c_t1 + c_t2 AS played, c_w AS won, CAST(c_w AS FLOAT) / (c_t1 + c_t2) * 100 AS won_percentage
    (SELECT team1, count(team1) AS c_t1 FROM ipl_m GROUP BY team1) AS t1 
    (SELECT team2, count(team2) AS c_t2 FROM ipl_m GROUP BY team2) AS t2 
ON t1.team1 = t2.team2
    (SELECT winner, count(winner) AS c_w FROM ipl_m GROUP BY winner) AS w
ON t1.team1 = w.winner OR t2.team2 = w.winner
ORDER BY won_percentage DESC;

Resulting Table:

team played won won_percentage
Chennai Super Kings 178 106 59.55056179775281
Mumbai Indians 203 120 59.11330049261084
Delhi Capitals 33 19 57.57575757575758
Sunrisers Hyderabad 124 66 53.2258064516129
Kolkata Knight Riders 192 99 51.5625

Table Definition:

    id integer PRIMARY KEY,
    match_id integer NOT NULL,
    city VARCHAR(20) NOT NULL,
    date DATE NOT NULL,
    player_of_match VARCHAR(50),
    venue VARCHAR(75) NOT NULL,
    neutral_venue BOOLEAN NOT NULL,
    team1 VARCHAR(50) NOT NULL,
    team2 VARCHAR(50) NOT NULL,
    toss_winner VARCHAR(50) NOT NULL,
    toss_decision VARCHAR(5) NOT NULL,
    winner VARCHAR(50),
    result VARCHAR(10),
    result_margin float,
    eliminator CHAR(1) NOT NULL,
    method VARCHAR(3),
    umpire1 VARCHAR(50),
    umpire2 VARCHAR(50)

1 Answer 1


Each row in ipl_m table has one winner and one loser. So first extract winners and set field result (it will be used in counting) to 1:

  winner AS team,
  1 as result
FROM ipl_m

Next extract losers and set field result to 0:

    WHEN team1 = winner THEN team2
    ELSE team1
  AS team,
  0 as result
FROM ipl_m

Combine two sets with UNION. Now SELECT from resulting set grouping by team column.

SELECT t.team AS team
, COUNT(*) AS played
, SUM(t.result) AS won
  winner AS team,
  1 as result
FROM ipl_m
    WHEN team1 = winner THEN team2
    ELSE team1
  AS team,
  0 as result
FROM ipl_m
) AS t
GROUP BY t.team

Your solution uses 4 SELECT and 2 JOIN operators. Mine uses 3 SELECT and 1 UNION. Using fewer operations is usually preferred.

  • \$\begingroup\$ Thanks for the answer. Initially your code didn't work as is, just needed to add an END to that CASE statement. Also UNION removes duplicates, so UNION ALL should be used instead. After making those changes, things work perfectly! \$\endgroup\$
    – gautham
    Nov 2, 2021 at 11:13
  • \$\begingroup\$ CASE ... END - my bad. UNION ALL - you are right ! \$\endgroup\$
    – JulStrat
    Nov 2, 2021 at 11:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.