t;[2]
> https://ci.apache.org/projects/flink/flink-docs-master/dev/stream/operators/process_function.html#example
>
>
> --Original Mail --
> *Sender:*aj
> *Send Date:*Fri May 29 02:07:33 2020
> *Recipients:*Yun Gao
> *CC:*user
> *Su
Window with multiple trigger condition
Hi,
I have implemented the below solution and its working fine but the biggest
problem with this is if no event coming for the user after 30 min then I am not
able to trigger because I am checking
time diff from upcoming events. So when the next event comes than
v/stream/operators/process_function.html#example
>
> Best,
> Yun
>
>
>
> --Original Mail --
> *Sender:*aj
> *Send Date:*Sun May 24 01:10:55 2020
> *Recipients:*Tzu-Li (Gordon) Tai
> *CC:*user
> *Subject:*Re: Flink Window with mult
v/stream/operators/process_function.html#example
Best,
Yun
--Original Mail --
Sender:aj
Send Date:Sun May 24 01:10:55 2020
Recipients:Tzu-Li (Gordon) Tai
CC:user
Subject:Re: Flink Window with multiple trigger condition
I am still not able to get much a
I am still not able to get much after reading the stuff. Please help with
some basic code to start to build this window and trigger.
Another option I am thinking is I just use a Richflatmap function and use
the keyed state to build this logic. Is that the correct approach?
On Fri, May 22, 2020
I was also thinking to have a processing time window but that will not work
for me. I want to start the window when the user "*search*" event arrives.
So for each user window will start from the *search* event.
The Tumbling window has fixed start end time so that will not be suitable
in my case.
Hi,
To achieve what you have in mind, I think what you have to do is to use a
processing time window of 30 mins, and have a custom trigger that matches
the "start" event in the `onElement` method and return
TriggerResult.FIRE_AND_PURGE.
That way, the window fires either when the processing time h
Session window defined on the gap of inactivity, I do not have that
requirement.
Start the window only on the "*search even*t" that part I will take later.
Let's say in the first phase I want to start the window on any event that
appears for that user.
For example :
*Scenario -1*
t1 - user1
Hello All,
I am getting a lot of user events in a stream. There are different types
of events, now I want to build some aggregation metrics for the user by
grouping events in buckets.
My condition for windowing is :
1. Start the window for the user when event_name: *"search" *arrived for
the u
