网站建设详细设计网络营销的策略

使用 com.baomidou.mybatisplus.mapper.BaseMapper#selectPage 方法查询分页列表时，在mysql层面会执行两条sql，一条是count，另一条才是获取分页数据，大部分场景下我们都可以直接使用这个方法来进行查询，但是一些特殊场景下这个获取总数的sql会比获取分页数据执行的耗时更高。

列子

1.count 查询

    SELECT user_identifier, transaction_date, monetary_unit, SUM(incoming_value) AS total_incoming, SUM(ABS(outgoing_value)) AS total_outgoing, SUM(closing_value) AS total_balance FROM financial_records_summary WHERE user_identifier = ? AND transaction_date >= ? AND transaction_date <= ? GROUP BY user_identifier, transaction_date, monetary_unit ORDER BY transaction_date DESC, monetary_unit, total_balance DESC
) AS aggregated_results

2.分页信息查询

SELECT user_identifier, transaction_date, monetary_unit, SUM(incoming_value) AS total_incoming, SUM(ABS(outgoing_value)) AS total_outgoing, SUM(closing_value) AS total_balance 
FROM financial_records_summary 
WHERE account_owner = ? AND transaction_date >= ? AND transaction_date <= ? 
GROUP BY user_identifier, transaction_date, monetary_unit 
ORDER BY transaction_date DESC, monetary_unit, total_balance DESC 
LIMIT 0, 10

当数据量足够大的时候，第一个sql查询耗时是第二个sql查询耗时的两倍多，原因主要是以下几点：

• COUNT需要处理全部结果集
• 数据查询有LIMIT限制可能提前终止
• 子查询中的ORDER BY对COUNT来说是多余开销

通过移除不必要的操作可以大幅度优化这个sql的性能