把表结构粘贴出来，如果能一点测试数据更好，不然不好调试改写。

1）要注意关键的问题，row是关键字，可以统一改成 row_
2）主要是 SELECT @row...@row:=-1) r ) se 这部分的改写，可以考虑统一换成如下内容，看看哪种是你需要的。

第一种：

with a as(select level-1 from dual connect by level<=10)
             select rownum row_,* from a a,a b,a c,a d

第二种：

select level row_ from dual connect by level<=10000

1. 关键字冲突处理将 ROW 替换成 row_num，避免与保留关键字冲突。
2. 去掉 MySQL 的 @变量 写法达梦不支持 @row 这种变量声明，改用 LEVEL 或 CONNECT BY 来生成序列。
3. 日期函数替换 达梦数据库用 TO_CHAR 替 DATE_FORMAT，用 ADD_MONTHS 替代 DATE_ADD。
4. 修改后的语句：

SELECT
A.date_time AS datetime,
COALESCE(B.num, 0) AS num,
B.title
FROM (
SELECT
TO_CHAR(ADD_MONTHS(DATE '2020-01-01', (row_num - 1) * 12), 'YYYY') AS date_time
FROM (
SELECT LEVEL AS row_num
FROM DUAL
CONNECT BY LEVEL <= 5
) t
) A
LEFT JOIN (
SELECT
TO_CHAR(PC_FSTIME, 'YYYY') AS date_time,
b.PC_RANK AS title,
COUNT(1) AS num
FROM qms_deviation_description a
JOIN qms_deviation_evaluation_classification b ON a.id = b.P_ID
WHERE b.PC_RANK IS NOT NULL
GROUP BY TO_CHAR(PC_FSTIME, 'YYYY'), b.PC_RANK
) B ON A.date_time = B.date_time;

根据你原逻辑同等改写

   SELECT A.date_time as datetime, 
          COALESCE ( B.num, 0 ) AS num, 
          B.title 
     FROM( 
		   SELECT DATE_FORMAT(dateadd(year, (rownum - 1), '2020-01-01'), '%Y' )  date_time
		   connect by rownum<=1e4
		   and DATE_FORMAT(dateadd(year, (rownum - 1), '2020-01-01'), '%Y' ) <= DATE_FORMAT((date'2024-12-01'), '%Y')
         ) A 
LEFT JOIN ( SELECT DATE_FORMAT( PC_FSTIME, '%Y' ) AS date_time, 
                   b.PC_RANK AS title, 
                   COUNT( 1 ) AS num 
              FROM qms_deviation_description a, 
                   qms_deviation_evaluation_classification b 
             WHERE a.id=b.P_ID 
               and b.PC_RANK IS NOT NULL 
          GROUP BY date_time, 
                   b.PC_RANK ) B 
       ON A.date_time = B.date_time