You are right, WindowFunctions collect all data in a window and are
evaluated at once.
Although FoldFunctions could be directly applied on each element that
enters a window, this is not done at the moment.
Only ReduceFunctions are eagerly applied.
If you port your code to a ReduceFunction, you can
I will give this a try.
Though I'm not sure I can switch over to WindowFunction.
I work with potentially huge Windows, the Fold gives me a minimal and
constant memory footprint. Switching to WindowFunction will require to keep
the Window in Memory before it can be processed (at least to my
underst
Sure. You don't need a trigger, but a WindowFunction instead of the
FoldFunction.
Only the WindowFunction has access to the Window object.
Something like this:
poissHostStreams
.timeWindow(Time.of(WINDOW_SIZE, TimeUnit.MILLISECONDS))
.apply(new WindowFunction() {
@overr
Hi Fabian,
thanks for your answer. Can I do the same in java using normal time windows
(without additional trigger)?
My current codes looks like this:
poissHostStreams
.timeWindow(Time.of(WINDOW_SIZE, TimeUnit.MILLISECONDS))
.fold(new Tuple2<>("", new HashMap<>()), new
MultiValue
Hi Martin,
you can get the start and end time of a window from the TimeWindow object.
The following Scala code snippet shows how to access the window end time
(start time is equivalent):
.timeWindow(Time.minutes(5))
.trigger(new EarlyCountTrigger(earlyCountThreshold))
.apply { (
key: Int,
win
Hej,
Is it possible to extract the start and end window time stamps from within
a window operator?
I have an event time based window that does a simple fold function. I want
to put the output into elasticsearch and want to preserve the start and end
timestamp of the data so I can directly compare
