Once you have a RelationalGroupedDataSet , you can use agg() to perform
group wide operation such max , sum , etc ... or even custom aggregator.
df.groupBy().agg(sum(col(...)))
That will return a DF with your groupBy columns and result of the
aggregation
Stephane
Soheil Pourbafrani wrote:
Hi,
Hi,
In my problem, I need to group the DataFrame, apply the business logic for
each group and finally emit a new DataFrame based on that. To describe in
detail, there is a device_dataframe which contains the timestamp of when
the device had been turned on (on) and turned off (off).
+-+---
