1.为了快速统计出支付宝的支付方式是否存在漏洞，对支付宝订单的回调建表存储
2.有以下三个表
select notify_id,out_trade_no,trade_no,COUNT(*),notify_time FROM alipay_callback WHERE trade_status = 'TRADE_SUCCESS' GROUP BY notify_id HAVING COUNT(*) > 1 ORDER BY notify_time
select notify_id,out_trade_no,trade_no,COUNT(*),notify_time FROM alipay_callback WHERE trade_status = 'TRADE_SUCCESS' GROUP BY out_trade_no HAVING COUNT(*) > 1 ORDER BY notify_time
select notify_id,out_trade_no,trade_no,COUNT(*),notify_time FROM alipay_callback WHERE trade_status = 'TRADE_SUCCESS' GROUP BY trade_no HAVING COUNT(*) > 1 ORDER BY notify_time
3.当上述三个表一致时视为数据正常，如何快速比较三个表，并且找出不一致的数据？
![[Web front-end] Node.js quick start](https://img.php.cn/upload/course/000/000/067/662b5d34ba7c0227.png)
Personally, I suggest you do it in your application. Don’t put very complex logic in SQL, even if the SQL only has a few lines and the program has dozens of lines. SQL is the place for storage and query, and the program is where the logic is done. It is easy to implement, easy to maintain, easy to expand, and easy for future generations to understand
If you use an app to do this, I believe there will be no problem. If the data is large, do it in batches according to notify_time