MYSQL-實現ORACLE- row_number() over(partition by ) 分組排序功能最佳化

今天看了篇文章，被幾個地方轉載，但是效能不太好，因為不能評論，所以把最佳化思路寫在這裡。

1.確定需求: 根據部門來分組,顯示各員工在部門裡按薪水排名名次.

2.來建立執行個體資料:

drop table if exists heyf_t10;

create table heyf_t10 (empid int ,deptid int ,salary decimal(10,2) );insert into heyf_t10 values(1,10,5500.00),(2,10,4500.00),(3,20,1900.00),(4,20,4800.00),(5,40,6500.00),(6,40,14500.00),(7,40,44500.00),(8,50,6500.00),

(9,50,7500.00);

3. http://www.kaishixue.com/mysql/14.html 文章中SQL的實現

select empid,deptid,salary,rank

from (select heyf_tmp.empid, heyf_tmp.deptid, heyf_tmp.salary, @rownum:=@rownum+1, if(@pdept=heyf_tmp.deptid,@rank:=@rank+1,@rank:=1) as rank, @pdept:=heyf_tmp.deptid from (select empid,deptid,salary from heyf_t10 order by deptid asc ,salary desc ) heyf_tmp, (select @rownum:=0, @pdept:=null,@rank:=0) a ) result;

執行計畫如下

idselect_typetabletypepossible_keyskeykey_lenrefrowsfilteredExtra
1PRIMARY<derived2>ALLNULLNULLNULLNULL9100.00
2DERIVED<derived4>systemNULLNULLNULLNULL1100.00
2DERIVED<derived3>ALLNULLNULLNULLNULL9100.00
4DERIVEDNULLNULLNULLNULLNULLNULLNULLNULLNo tables used
3DERIVEDheyf_t10ALLNULLNULLNULLNULL9100.00Using filesort

如果大資料量直接就掛了。換個思路來寫

select h.`empid`,h.`deptid`,h.`salary`,count(*) as rank
from heyf_t10 as h
left outer join heyf_t10 as r
on h.deptid = r.deptid
and h.`salary`<= r.`salary`
group by h.`empid`,h.`deptid`,h.`salary`
order by h.deptid, h.salary desc;

idselect_typetabletypepossible_keyskeykey_lenrefrowsfilteredExtra
1SIMPLEhALLNULLNULLNULLNULL9100.00Using temporary; Using filesort
1SIMPLErALLNULLNULLNULLNULL9100.00

需要最佳化的時候建立個索引最佳化掉filesort，很直觀，效率也提高了很多。有時候不能太趕需求，遇到複雜的SQL還是讓DBA來最佳化一下。

--EOF--