flink 消费kafka 数据去重后聚合问题

[yanyunpeng]

create view distinct_view as 
select val,ptime from (select * ,ROW_NUMBER() OVER (PARTITION BY MN ORDER BY 
ptime ) as rowNum from test_udf) where rowNum = 1


select avg(val) as avg_var, STDDEV_POP(val) as spddev_pop_var from 
distinct_view GROUP BY HOP(ptime, INTERVAL '2' SECONDS, INTERVAL '1' DAY);

kafka的数据有重复 但是这么写有问题 请问这种情况应该怎么写