2、shardingjdbc top n工作原理分析

shardingjdbc在当前分库分表领域占有很大市场份额。可能和其宣传有关系，其文档突出自己的优点，对自己的缺点写的很少或者很不容易被注意到。所以，我们在使用shardingjdbc做业务的时候要关注他最终生成的sql是什么？是否符合我们的业务需求。这里不是黑shardingjdbc，不是shardingjdbc的问题，这是计算机行业难以跨越的问题。

我们就拿一份按照季度的分库分表场景做演示吧。这里可以看出shardingjdbc取了intmax的分表记录，然后汇聚取top10.可见有多么大的网络开销。所以涉及排序的场景要做好技术测试。

select  station_id,count(station_id) as stationNum  from tms_delivery_order_2021   
where delivery_start_time between  '2021-01-01 01:41:36' and  '2021-12-30 01:41:36'  
group by  station_id  order by stationNum desc LIMIT 10

#sql1 第一季度
select  station_id,count(station_id) as stationNum  from tms_delivery_order_2021_q1   
where delivery_start_time between  '2021-01-01 01:41:36' and  '2021-12-30 01:41:36'  
group by  station_id  order by stationNum desc LIMIT 2147483647

#sql2 第二季度
select  station_id,count(station_id) as stationNum  from tms_delivery_order_2021_q2   
where delivery_start_time between  '2021-01-01 01:41:36' and  '2021-12-30 01:41:36'  
group by  station_id  order by stationNum desc LIMIT 2147483647

#sql3 第三季度
select  station_id,count(station_id) as stationNum  from tms_delivery_order_2021_q3   
where delivery_start_time between  '2021-01-01 01:41:36' and  '2021-12-30 01:41:36'  
group by  station_id  order by stationNum desc LIMIT 2147483647

#sql1 第四季度
select  station_id,count(station_id) as stationNum  from tms_delivery_order_2021_q4   
where delivery_start_time between  '2021-01-01 01:41:36' and  '2021-12-30 01:41:36'  
group by  station_id  order by stationNum desc LIMIT 2147483647