php - SQL LEFT JOIN 不使用主键的查询太慢，如何加快速度？

Question

我有三个 MySQL 表； “teams”和两个表，具体取决于团队类型(“team_type_1”和“team_type_2”)。我需要将所有三个表查询到一个表中，并将两个 team_type 表中每个团队的英里数汇总到一个“英里”列中。

下面的查询可以根据需要工作，但由于我的 JOIN 查询不使用主 ID，它要么锁定，要么花费太长时间。

在 JOIN 中使用主 ID“team_type_1.id”和“team_type_2.id”会很好，但我在下面的最后一个查询中的尝试仍然在内部查询中使用非主键。我可以采取不同的方法来加快查询速度吗？

Table 1: teams (3000 total records)
id (primary) | team_name


Table 2: team_type_1 (12000 total records)
id (primary) | team_id | miles


Table 3: team_type_2 (150000 total records)
id (primary) | team_id | miles

---

将上表合并为: 团队名称 |英里

---

当使用主键单独运行这些查询时，只需要几毫秒，但我需要组合这些表:

SELECT
teams.team_name,
SUM(team_type_1.miles)
FROM team_type_1
LEFT JOIN teams ON teams.id = team_type_1.team_id
GROUP BY team_type_1.team_id


SELECT
teams.team_name,
SUM(team_type_2.miles)
FROM team_type_2
LEFT JOIN teams ON teams.id = team_type_2.team_id
GROUP BY team_type_2.team_id

---

当组合且不使用主键时，它会锁定数据库:

SELECT
teams.team_name,
SUM(CASE WHEN team_type_2.miles THEN team_type_2.miles ELSE team_type_1.miles END)
FROM teams

LEFT JOIN team_type_1 ON team_type_1.team_id = teams.id
LEFT JOIN team_type_2 ON team_type_2.team_id = teams.id

GROUP BY teams.id

---

运行大约需要 8-10 秒，但内部选择使用非主 ID；有没有办法加快速度？

SELECT
teams.team_name,
SUM(CASE WHEN team_type_2.miles THEN team_type_2.miles ELSE team_type_1.miles END)
FROM teams

LEFT JOIN team_type_1 ON team_type_1.id = (SELECT team_type_1.id FROM team_type_1 WHERE team_type_1.team_id = teams.id GROUP BY team_type_1.team_id)
LEFT JOIN team_type_2 ON team_type_2.id = (SELECT team_type_2.id FROM team_type_2 WHERE team_type_2.team_id = teams.id GROUP BY team_type_2.team_id)

WHERE team_type_2.miles > 0 OR team_type_1.miles > 0 GROUP BY teams.id

---

• 更新:

感谢@Strawberry 的 UNION 提示，除了上面的几列和三个表之外，我还有正在查询的其他表和列，但没有提及这些，因为我可以使用主 ID对于那些。下面是我使用 UNIONS 的完整查询结构，它正在按我的需要工作。

SELECT
teams.team_name,
teams.start_date,
team_counties.county_name,
team_leagues.league_name,
SUM(team_type_1.miles)
FROM team_type_1

LEFT JOIN teams ON teams.id = team_type_1.team_id
LEFT JOIN team_counties ON team_counties.id = teams.county_id
LEFT JOIN team_leagues ON team_leagues.id = teams.league_id

GROUP BY team_type_1.team_id

UNION

SELECT
teams.team_name,
teams.start_date,
team_counties.county_name,
team_leagues.league_name,
SUM(team_type_2.miles)
FROM team_type_2

LEFT JOIN teams ON teams.id = team_type_2.team_id
LEFT JOIN team_counties ON team_counties.id = teams.county_id
LEFT JOIN team_leagues ON team_leagues.id = teams.league_id

GROUP BY team_type_2.team_id

Answer 1

有几种方法可以解决这个问题，但最简单的解决方案就是简单地 index两个子表中的 team_id 列:

ALTER TABLE `team_type_1`
    ADD INDEX `team_id` (`team_id`);
ALTER TABLE `team_type_2`
    ADD INDEX `team_id` (`team_id`);

这应该可以让您以一定的速度执行以下操作:

SELECT teams.id,
       teams.team_name,
       SUM(IFNULL(team_type_1.miles, 0) + IFNULL(team_type_2.miles, 0)) sum_miles
FROM teams
LEFT JOIN team_type_1 ON team_type_1.team_id = teams.id
LEFT JOIN team_type_2 ON team_type_2.team_id = teams.id
GROUP BY teams.id,
         teams.team_name