这个问题可以用SQL 实现吗?

分别统计2次,就可以了.
ID,Date,Comment
1,04/05/2017,''
1,04/05/2017,''
1,04/05/2017,''
2,05/05/2017,''
2,06/05/2017,''
2,07/05/2017,''
2,08/05/2017,''
2,05/05/2017,''
select id,date,count(1) as cnt from 表 group by id,date
-- 结果
id,date,cnt
1,04/05/2017,3
2,05/05/2017,2
2,06/05/2017,1
2,07/05/2017,1
2,08/05/2017,1
select id,count(1) as cnt2 from 表 group by id
-- 结果
id,date,cnt2
1,1
2,4
把上面的2步组合起来,
update 表 set
from 表 a,(select id,coment =(case when b.cunt >=3 then 'yes' else 'no' end) from (select id,count(1) as cnt2 from (select id,date,count(1) as cnt from 表 group by id,date) group by id)b)c
where a.id=c.id