减少子查询的关键是把嵌套逻辑“拉平”，用连接（JOIN）或临时表替代，让MySQL执行计划更高效、更可控。
用JOIN代替相关子查询
相关子查询（即子查询依赖外层表字段）每行执行一次，性能极差。多数场景可改写为LEFT JOIN或INNER JOIN。
原写法：SELECT id, name FROM users WHERE id IN (SELECT user_id FROM orders WHERE status = 'paid') 优化后：SELECT DISTINCT u.id, u.name FROM users u INNER JOIN orders o ON u.id = o.user_id WHERE o.status = 'paid' 注意：用DISTINCT防重复；若需保留无订单用户，改用LEFT JOIN + WHERE o.user_id IS NOT NULL用派生表（Derived Table）替代非相关子查询
不依赖外层的子查询（如聚合结果），可提前算好作为临时结果集，避免多次计算。
原写法：SELECT name, (SELECT AVG(score) FROM exams) AS avg_score FROM students 优化后：SELECT s.name, dt.avg_score FROM students s CROSS JOIN (SELECT AVG(score) AS avg_score FROM exams) dt CROSS JOIN更语义清晰；也可用JOIN（无ON条件）或直接在SELECT中用变量（需确保单行结果）用EXISTS替代IN（尤其大表子查询）
当子查询返回大量ID，且只需判断存在性时，EXISTS通常比IN更快，因为它能短路退出。
原写法：SELECT * FROM products WHERE category_id IN (SELECT id FROM categories WHERE active = 1) 优化后：SELECT * FROM products p WHERE EXISTS (SELECT 1 FROM categories c WHERE c.id = p.category_id AND c.active = 1) 确保子查询中关联字段有索引（如categories.id、products.category_id）必要时拆分+临时表缓存中间结果
复杂多层嵌套子查询，尤其是含GROUP BY、窗口函数或重复计算的，可先存入临时表，再JOIN使用。
例如统计每个用户最近3笔订单总金额，可先建临时表：CREATE TEMPORARY TABLE tmp_last3 AS SELECT user_id, SUM(amount) sum_amt FROM (SELECT user_id, amount, ROW_NUMBER() OVER (PARTITION BY user_id ORDER BY created_at DESC) rn FROM orders) t WHERE rn 再与users表JOIN：SELECT u.name, t.sum_amt FROM users u LEFT JOIN tmp_last3 t ON u.id = t.user_id 临时表自动索引主键，支持WHERE/JOIN加速，也便于EXPLAIN分析
