首先过滤每个employee_id的行，以获得所需的status_id。
然后，通过应用所需的优先级，使用ROW_NUMBER()窗口函数对行进行排名。
最后从每个排名中选择第一行：

WITH cte AS (
  SELECT *,
         ROW_NUMBER() OVER (
           PARTITION BY employee_id
           ORDER BY status_id = 7 DESC,
                    status_id = 4 DESC,
                    status_id = 3 DESC, -- this is the last option and it is actually not needed
                    id
         ) AS rn 
  FROM employee_leaves 
  WHERE status_id IN (7, 4, 3)
)
SELECT id, employee_id, status_id 
FROM cte 
WHERE rn = 1;

ROW_NUMBER()中的ORDER BY子句可以通过使用函数FIELD()来简化：

ORDER BY FIELD(status_id, 7, 4, 3),
         id

或者函数FIND_IN_SET()：

ORDER BY FIND_IN_SET(status_id, '7,4,3'),
         id

请参见demo。