GROUP BY有种合并同类项的感觉,其实也就是那个效果,我在这次的POSTER里写了一串,代码如下:

SELECT pr_rates.pid,count(pr_rates.ip),pr_works.title,pr_works.realname,pr_college.name FROM pr_works RIGHT JOIN pr_rates ON pr_rates.pid=pr_works.pid LEFT JOIN pr_college ON pr_works.college=pr_college.id GROUP BY pid ORDER BY count(pr_rates.ip) DESC

其中就用到了GROUP BY pid 相当于把相同的pid合并,然后用count()来统计就得出了得票数.