Technology sharing | a clustering incremental statistical SQL requirement in MySQL

author ： Liu Chen

Net name bisal , More than 10 years of experience in application operation and maintenance , At present, he is mainly engaged in the work related to the improvement of database application R & D capability and technical management ,Oracle ACE , Tencent cloud TVP, Have Oracle OCM & OCP 、EXIN DevOps Master 、SCJP And other international certification , The first batch in China Oracle YEP member ,OCMU member ,《DevOps Best practices 》 One of the Chinese translators ,CSDN & ITPub Expert bloggers , official account "bisal My personal grocery store ", Sharing technical articles for a long time , Share technical topics online and offline many times .

In this paper, the source ： Original contribution

* Produced by aikesheng open source community , Original content is not allowed to be used without authorization , For reprint, please contact the editor and indicate the source .

---

My colleague raised a question MySQL In the database SQL The problem of statistics , I use test data to simulate , As shown below , surface tt There are three fields ,code Is the identification name ,cdate Is the corresponding date ,ctotal It's a statistic

The original statistical statement is as follows , according to code and cdate Clustering , Count out each code The number of records accumulated each month ,

But his need is to be able to follow code+cdate The statistical value of is displayed accumulatively , For example, the first line in the above figure total yes 10 , Second elements total Show 10+9=19 , In the third line total Show 10+9+11=30 , And so on .

What he needs is total Line by line ,MySQL Can be achieved by defining variable accumulation , As shown below , utilize @i Line by line ,

It really looks like it has come true , But there's a problem , Different code , No zeroing is done , For example, in the figure above ,code=BBB The first line of the record ,total Should be 5 , Current 35 Is obviously through 30+5 Got .

There should be another logic , Namely, in accordance with the code While clustering sum Sum calculation , As shown below , Now you can see ,g_total according to code and cdate Add up , Different code , It will return to zero , Recalculate , Meet the actual needs ,

In fact, this one can be optimized ,MySQL 8.0 Support with , Above SQL Chinese vs tt The table was read twice , With the help of with , It can be reduced to read only once tt surface ,

in addition , Combined with the actual retrieval scenario requirements , Consider adding indexes to related fields , Further improve the efficiency of data retrieval .

Of course , The above SQL There may be other alternatives or better solutions , This article only gives one of the solutions .

therefore , adopt SQL Achieve business needs , On the one hand, we need to fully understand the meaning of requirements , Can accurately map to specific SQL logically , On the other hand, you should know the functions supported by the database 、 function 、 Characteristics, etc. , Whether there is an implementation that can meet this requirement , besides , Non functional factors , It can't be ignored , The right index 、 Avoid reading data repeatedly 、 Avoiding unnecessary sorting and so on are the means we can use .