Problems about pv uv in flink sql

Posted by Joshua Fan on
URL: http://deprecated-apache-flink-user-mailing-list-archive.369.s1.nabble.com/Flink-streaming-sql-group-by-tp34412p40811.html

Hi 

I have learned from the community on how to do pv/uv in flink sql. One is to make a yyyyMMdd grouping, the other is to make a day window. Thank you all.

I have a question about the result output. For yyyyMMdd grouping, every minute the database would get a record, and many records would be in the database as time goes on, but there would be only a few records in the database according to the day window.

for example, the pv would be 12:00,100   12:01,200  12:02,300   12:03,400 according to the yyyyMMdd grouping solution, for the day window solution, there would be only one record as  12:00,100 |12:01,200|12:02,300|12:03,400.

I wonder, for the day window solution, is it possible to have the same result output as the yyyyMMdd solution? because the day window solution has no worry about the state retention.

Thanks.

Yours sincerely

Josh