select t1.mk,t2.gnjy,t1.rjcw,t3.total,count(t1.rjcw)/(t3.total*1.0) percentagefrom (select mk,count(cswhdxs) as rjcw from YSDP_DATA where YEAR(TJSJ)=2015 and CSWHDXS='軟件錯(cuò)誤' group by mk) t1 left join     (select mk,count(cswhdxs) as gnjy from YSDP_DATA where YEAR(TJSJ)=2015 and CSWHDXS='功能建議' group by mk) t2       on t1.mk=t2.mk left join     (SELECT mk,count(cswhdxs) AS total FROM YSDP_DATA where year(TJSJ)=2015 group by MK) t3      on t1.mk=t3.mkgroup by t1.MK,t1.rjcw,t2.gnjy ,t3.total 查詢語(yǔ)句如上，目的很簡(jiǎn)單，就是統(tǒng)計(jì)表中字段CSWHDXS中（特定年份）軟件錯(cuò)誤 和 功能建議 的數(shù)量并算出 軟件錯(cuò)誤 在該列的占比，最后用另一字段mk進(jìn)行分類(lèi)?，F(xiàn)在前4列應(yīng)該都沒(méi)有問(wèn)題，最后一列計(jì)算卻有明顯錯(cuò)誤，結(jié)果如下圖最后一列明顯計(jì)算錯(cuò)誤，求教大家原因在哪里，應(yīng)該如何修改。或者是我的思路根本不對(duì)。