优化查询速度 - SQL替换Service

Service 接口问题

下面是原先的 Service 实现类代码，有门店 ID、订单状态、查询时间段，然后查出了所有的结果，继续使用 java8 的特性获取汇总结果，随着项目的推移，数据量越来越大，这种方式是不可取的，随便都能上几十秒查不出来，导致请求超时，下面我们就来优化一下这个方法
在这里插入图片描述

SQL 优化

先把上面的方法都转换成 SQL 语句的写法，得到下面的 5 个查询 SQL

select sum(A.pay_money) originPrice, count(A.id) countOrder, sum(A.pay_money) payMoney
from biz_order_info A
where A.`status` = 1 and A.order_status = 2 and A.store_id = 1;

select sum(A.pay_money) smPayMoney
from biz_order_info A
where A.`status` = 1 and A.order_status = 2 and A.store_id = 1 and A.pay_way = 1;

select sum(A.pay_money) rlPayMoney
from biz_order_info A
where A.`status` = 1 and A.order_status = 2 and A.store_id = 1 and A.pay_way = 2;

select sum(A.pay_money) sjPayMoney
from biz_order_info A
where A.`status` = 1 and A.order_status = 2 and A.store_id = 1 and A.pay_way = 3;

select sum(A.pay_money) xjPayMoney
from biz_order_info A
where A.`status` = 1 and A.order_status = 2 and A.store_id = 1 and A.pay_way = 4;

整理一下代码，通过union all连接，最简单的方法，后端使用 List 来接收返回值，有一定的可行性

select * from (
	(select sum(A.pay_money) 'value' from biz_order_info A where A.`status` = 1 and A.order_status = 2 and A.store_id = 1)
	union all
	(select count(A.id) 'value' from biz_order_info A where A.`status` = 1 and A.order_status = 2 and A.store_id = 1)
	union all
	(select sum(A.pay_money) 'value' from biz_order_info A where A.`status` = 1 and A.order_status = 2 and A.store_id = 1)
	union all
	(select sum(A.pay_money) 'value' from biz_order_info A where A.`status` = 1 and A.order_status = 2 and A.store_id = 1 and A.pay_way = 1)
	union all
	(select sum(A.pay_money) 'value' from biz_order_info A where A.`status` = 1 and A.order_status = 2 and A.store_id = 1 and A.pay_way = 2)
	union all
	(select sum(A.pay_money) 'value' from biz_order_info A where A.`status` = 1 and A.order_status = 2 and A.store_id = 1 and A.pay_way = 3)
	union all
	(select sum(A.pay_money) 'value' from biz_order_info A where A.`status` = 1 and A.order_status = 2 and A.store_id = 1 and A.pay_way = 4)
) X;

查询结果
在这里插入图片描述
当我放到项目中时，发现还有查询条件，时间段，如果上面的 SQL 加上时间段，将会变的非常臃肿且难维护，所以上面的 SQL 报废了，不能使用，继续优化 SQL

-- 最开始的SQL
select A.*
from biz_order_info A
where A.`status` = 1 and A.order_status = 2;

-- 初步汇总，拿到3个我们需要的值
select sum(A.pay_money), count(A.id), sum(A.pay_money)
from biz_order_info A
where A.`status` = 1 and A.order_status = 2;

问题来了，怎么在已经查询出的结果中分支付方式汇总数据？

这里我使用的是 sum + case when 结构，分类汇总数据，再次优化 SQL

select sum(A.pay_money) originPrice, count(A.id) countOrder, sum(A.pay_money) payMoney,
sum(case A.pay_way when '1' THEN A.pay_money ELSE 0 END) smPayMoney,
sum(case A.pay_way when '2' THEN A.pay_money ELSE 0 END) rlPayMoney,
sum(case A.pay_way when '3' THEN A.pay_money ELSE 0 END) sjPayMoney,
sum(case A.pay_way when '4' THEN A.pay_money ELSE 0 END) xjPayMoney
from biz_order_info A
where A.`status` = 1 and A.order_status = 2