Apache Flink

:wave: I want to dedupe my stream and keep records with maximum processedCount. Would the following be correct to use if I need to do event time processing afterwards?
```SELECT *
FROM (
  SELECT *,
    ROW_NUMBER() OVER (PARTITION BY id ORDER BY processedCount DESC) AS row_num
  FROM event)
WHERE row_num = 1```
note `ORDER BY processedCount` . I know it does not translate to Deduplicate optimized plan because I am not using time attribute as the <https://nightlies.apache.org/flink/flink-docs-release-1.15/docs/dev/table/sql/queries/deduplication/|Flink doc> suggests. My question is that whether the result preserve the watermark?

Originally, I was using a regular self-join but then came across this <https://stackoverflow.com/questions/74971028/apache-flink-issue-with-joins-on-kinesis-streams-rowtime-attributes-must-not-be/74978168#74978168|stackoverflow question>
&gt; It is not possible for a regular join to have time attributes in its results, because the time attribute could not be well-defined.
I am wondering if my dedupe query above will have some kind of similar limitation.