如何在 SQLite 中减去 2 列？答案

【问题标题】：how to Subtract 2 Columns in SQLite?如何在 SQLite 中减去 2 列？
【发布时间】：2018-02-14 23:09:37
【问题描述】：

如果给定表格：

games_played(ID,home_team,away_team,homescore,awayscore)

输入：

('1', 'USA', 'CAN', '0','2')

('2', 'USA', 'CAN', '1','2')

('3', 'USA', 'CAN', '3','0')

('4', 'SWE', 'CAN', '3','2')

('5', 'CAN', 'RUS', '3','2')

('6', 'DE', 'RUS', '0','2')

输出应该是团队：SWE、CAN

我会怎么做才能找到主场胜负多的球队？我试过这样做：

select home_team, count(*)
  from games_played
  where homescore > awayscore
  group by home_team

select home_team, count(*)
  from games_played
  where awayscore > homescore
  group by home_team

第一个查询应该返回球队名称和他们在主场赢得的比赛场数。第二个应该是他们在主场输掉的比赛场数。此查询不起作用，我正在寻找另一种方法。我知道我不能在 'where' 中使用 count

【问题讨论】：

好的，我已经编辑了输入和输出。

标签： sql sqlite

【解决方案1】：

使用条件聚合：

SELECT home_team,
COUNT(CASE WHEN homescore > awayscore THEN 1 END) AS gamesWon,
COUNT(CASE WHEN awayscore > homescore THEN 1 END) AS gamesLost
FROM games_played
GROUP BY home_team

结果：

现在只返回将gamesWon > gamesLost 放入common table expression 的团队：

WITH cte AS (
SELECT home_team,
COUNT(CASE WHEN homescore > awayscore THEN 1 END) AS gamesWon,
COUNT(CASE WHEN awayscore > homescore THEN 1 END) AS gamesLost
FROM games_played
GROUP BY home_team)

SELECT * FROM cte
WHERE gamesWon > gamesLost

结果：

复制：

[DEMO HERE]

CREATE TABLE games_played (ID INT,home_team VARCHAR(20),away_team VARCHAR(20),homescore INT,awayscore INT);

/* Create few records in this table */
INSERT INTO games_played VALUES(1,'USA','CAN',0,2);
INSERT INTO games_played VALUES(2,'USA','CAN',1,2);
INSERT INTO games_played VALUES(3,'USA','CAN',3,0);
INSERT INTO games_played VALUES(4,'SWE','CAN',3,2);
INSERT INTO games_played VALUES(5,'CAN','RUS',3,2);
INSERT INTO games_played VALUES(6,'DE','RUS',0,2);


/* Verify */
-- SELECT * FROM games_played;

WITH cte AS (
SELECT home_team,
COUNT(CASE WHEN homescore > awayscore THEN 1 END) AS gamesWon,
COUNT(CASE WHEN awayscore > homescore THEN 1 END) AS gamesLost
FROM games_played
GROUP BY home_team)

SELECT * FROM cte
WHERE gamesWon > gamesLost

【讨论】：

我将如何进一步使用它来隔离 CAN 和 SWE 国家，因为它们是仅有的 2 个主场赢比输多的国家？而不是输出所有主场胜负的计数。我想我需要在 home_win > home_loss 时创建视图并输出？有没有办法不使用创建视图
@s878 是的，请查看我的最终编辑以及解释和演示。

【解决方案2】：

这是一个更简单的方法，

SELECT g.home_team AS Team
FROM games_played g
LEFT JOIN games_played Win ON Win.id = g.id AND Win.homescore > Win.awayscore
LEFT JOIN games_played Lost ON Lost.id = g.id AND Lost.homescore < Lost.awayscore
GROUP BY g.home_team
HAVING
COUNT(Win.home_team) > COUNT(Lost.home_team)

【讨论】：

这不会只为这种情况输出正确答案，但如果有一个实例，例如我们有 (1,'USA','CAN',3,2) 那么它不会'行不通，因为美国会输掉 1 场主场比赛，但主场比赛赢得的比赛数将超过输球，应该输出。