多表级联查询分组函数慢问题

4张表级联查询的结果分组求和:

select sum(c.quantity) as quantity,sum(c.prefee) as fee
from  Cs_AccountBill a
join Cs_AccountBillDetail b on a.id= b.AccountBillId
join Cs_SettlementDetail c on c.accountBillDetailID = b.id 
join Cs_Settlement d on c.settlementID = d.id

While it's about 21W records of the result，time of select is about 31 ms.

更改如下：

select c.quantity,c.prefee
from  Cs_AccountBill a
join Cs_AccountBillDetail b on a.id= b.AccountBillId
join Cs_SettlementDetail c on c.accountBillDetailID = b.id 
join Cs_Settlement d on c.settlementID = d.id

　　

查询时间：40ms左右

再将查询的结果map reduce 大概80+ms

BigDecimal sum = list.parallelStream().map(Person::getFee).reduce(BigDecimal::add).get();

//勿用list.stream(),慢