优化 mysql 商品销售情况统计查询性能
问题描述：
针对如下 sql 查询：
select p.title, count(o.id) as total, coalesce(sum(o.amount), 0) as success_amount, coalesce(sum(success.amount), 0) as failed_amount, coalesce(sum(failed.amount), 0) as total_amount, count(success.id) as success_total, count(failed.id) as failed_total from goods as g left join orders as o on o.goods_id = g.id left join orders as success on success.goods_id = g.id and success.status = 1 left join orders as failed on failed.goods_id = g.id and failed.status = 2 group by `p`.`id` order by total desc limit 10
当查询时间范围较小时，查询速度较快，但当查询范围扩大时，查询速度极慢。
优化建议：
- 取消索引：删除 goods 表的 create_time 索引。
- 修改索引：将 orders 表的 goods_id 索引修改为 (create_time, goods_id, amount, status) 联合索引。
- 重写 sql：优化后的 sql 如下：
SELECT g.title, COUNT(*) AS total, COALESCE(SUM(o.amount), 0) AS total_amount, COALESCE(SUM(IF(o.status = 1, o.amount, 0)), 0) AS success_amount, COALESCE(SUM(IF(o.status = 2, o.amount, 0)), 0) AS failed_amount, COALESCE(SUM(o.status = 1), 0) AS success_total, COALESCE(SUM(o.status = 2), 0) AS failed_total FROM orders AS o JOIN goods AS g ON g.id = o.goods_id WHERE o.create_time BETWEEN 'xxx' AND 'yyy' GROUP BY o.id ORDER BY total DESC LIMIT 10
注：
由于数据量较小（商品数量：8000，订单数量：100000），考虑使用 sqlite 代替 mysql 进行查询，可以获得较好的性能。
© 版权声明
文章版权归作者所有，未经允许请勿转载。
THE END
