I have a use case where I would like to compute windows with pre-determined start times and end times. For example, I want to compute all events within April 1st at 1pm to April 7th at 1pm.
I figured I can compute daily windows and keep the results in state, but this seems optimal.
Does Flink support time windowing at pre-determined start and end times?
Seems like you could implement this as either a batch job or as a bounded streaming job, and simply aggregate all of the data together, not with a window, but with reduce or something similar.