(DEPRECATED) Apache Flink User Mailing List archive.

Apache Flink - Question about dynamically changing window end time at run time

classic Classic list List threaded Threaded
♦
♦
 8 messages Options
M Singh
Reply | Threaded
Open this post in threaded view
♦
♦
|

Apache Flink - Question about dynamically changing window end time at run time

M Singh
Hi:

I am working on a project and need to change the end time of the window dynamically.  I want to find out if the end time of the window is used internally (for sorting windows/etc) except for handling watermarks that would cause problems if the end time was changed during run time after the window has been created even if no new event has arrived for that window.

I don't want to use GlobalWindow since from my understanding it never gets destroyed.

If there is any alternate way of dealing with this, please let me know.

Thanks

Mans
Rong Rong
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Apache Flink - Question about dynamically changing window end time at run time

Rong Rong
Hi Mans,

I am not sure what you meant by "dynamically change the end-time of a window. If you are referring to dynamically determines the firing time of the window, then it fits into the description of session window [1]: 
If you want to handle window end time dynamically, one way of which I can think of is the dynamic gap, session window [1] approach. with which you can specify the end-time of a window based on input elements. Provided that you are maintaining a session window. 
Another way to look at it is through the Flink-CEP library [2]. 

Thanks,
Rong



On Tue, Apr 23, 2019 at 8:19 AM M Singh <[hidden email]> wrote:
Hi:

I am working on a project and need to change the end time of the window dynamically.  I want to find out if the end time of the window is used internally (for sorting windows/etc) except for handling watermarks that would cause problems if the end time was changed during run time after the window has been created even if no new event has arrived for that window.

I don't want to use GlobalWindow since from my understanding it never gets destroyed.

If there is any alternate way of dealing with this, please let me know.

Thanks

Mans
M Singh
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Apache Flink - Question about dynamically changing window end time at run time

M Singh
Hi Rong:

Thanks for your answer.

From what I understand the dynamic gap session windows are also created when the event is encountered.  I need to be able to change the window end time at a later time based on what other events are in that window.  One way to do this is to use GlobalWindows but then these are never deleted.

Regarding CEP option - I believe that CEP patterns cannot be changed dynamically once they've been complied which limits it usage.

Please feel free to correct me.

Thanks for your help and pointers.

On Tuesday, April 23, 2019, 8:12:56 PM EDT, Rong Rong <[hidden email]> wrote:


Hi Mans,

I am not sure what you meant by "dynamically change the end-time of a window. If you are referring to dynamically determines the firing time of the window, then it fits into the description of session window [1]: 
If you want to handle window end time dynamically, one way of which I can think of is the dynamic gap, session window [1] approach. with which you can specify the end-time of a window based on input elements. Provided that you are maintaining a session window. 
Another way to look at it is through the Flink-CEP library [2]. 

Thanks,
Rong



On Tue, Apr 23, 2019 at 8:19 AM M Singh <[hidden email]> wrote:
Hi:

I am working on a project and need to change the end time of the window dynamically.  I want to find out if the end time of the window is used internally (for sorting windows/etc) except for handling watermarks that would cause problems if the end time was changed during run time after the window has been created even if no new event has arrived for that window.

I don't want to use GlobalWindow since from my understanding it never gets destroyed.

If there is any alternate way of dealing with this, please let me know.

Thanks

Mans
Sameer Wadkar
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Apache Flink - Question about dynamically changing window end time at run time

Sameer Wadkar
Global Windows is fine for this use case. I have used the same strategy. You just define custom evictors and triggers and you are all good. Windows are managed by keys, so as such as long as events are evicted from the window, that counts towards reclaiming memory for the key+window combination. Plus there is just window per key with Global Windows. 

On Wed, Apr 24, 2019 at 7:47 AM M Singh <[hidden email]> wrote:
Hi Rong:

Thanks for your answer.

From what I understand the dynamic gap session windows are also created when the event is encountered.  I need to be able to change the window end time at a later time based on what other events are in that window.  One way to do this is to use GlobalWindows but then these are never deleted.

Regarding CEP option - I believe that CEP patterns cannot be changed dynamically once they've been complied which limits it usage.

Please feel free to correct me.

Thanks for your help and pointers.

On Tuesday, April 23, 2019, 8:12:56 PM EDT, Rong Rong <[hidden email]> wrote:


Hi Mans,

I am not sure what you meant by "dynamically change the end-time of a window. If you are referring to dynamically determines the firing time of the window, then it fits into the description of session window [1]: 
If you want to handle window end time dynamically, one way of which I can think of is the dynamic gap, session window [1] approach. with which you can specify the end-time of a window based on input elements. Provided that you are maintaining a session window. 
Another way to look at it is through the Flink-CEP library [2]. 

Thanks,
Rong



On Tue, Apr 23, 2019 at 8:19 AM M Singh <[hidden email]> wrote:
Hi:

I am working on a project and need to change the end time of the window dynamically.  I want to find out if the end time of the window is used internally (for sorting windows/etc) except for handling watermarks that would cause problems if the end time was changed during run time after the window has been created even if no new event has arrived for that window.

I don't want to use GlobalWindow since from my understanding it never gets destroyed.

If there is any alternate way of dealing with this, please let me know.

Thanks

Mans
Rong Rong
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Apache Flink - Question about dynamically changing window end time at run time

Rong Rong
Hi Mans,

Sameer is correct. if you would like to control window triggering based on other elements that does not belong to this window (in a keyed stream context) then this is probably the best way to approach. 

I think you've also posted in another thread that describes what will be left after fire-and-purge [1]. As Fabian stated: the only thing that might've left after is the window (which is the 2 long values indicate the start/end) and the trigger object. But you are right it might eventually filled up memory.

Another approach is to implement your own operator that handles all these internally by your user code. This would require you to replicate many of the window operator logic though.

Thanks,
Rong


On Wed, Apr 24, 2019 at 5:02 AM Sameer W <[hidden email]> wrote:
Global Windows is fine for this use case. I have used the same strategy. You just define custom evictors and triggers and you are all good. Windows are managed by keys, so as such as long as events are evicted from the window, that counts towards reclaiming memory for the key+window combination. Plus there is just window per key with Global Windows. 

On Wed, Apr 24, 2019 at 7:47 AM M Singh <[hidden email]> wrote:
Hi Rong:

Thanks for your answer.

From what I understand the dynamic gap session windows are also created when the event is encountered.  I need to be able to change the window end time at a later time based on what other events are in that window.  One way to do this is to use GlobalWindows but then these are never deleted.

Regarding CEP option - I believe that CEP patterns cannot be changed dynamically once they've been complied which limits it usage.

Please feel free to correct me.

Thanks for your help and pointers.

On Tuesday, April 23, 2019, 8:12:56 PM EDT, Rong Rong <[hidden email]> wrote:


Hi Mans,

I am not sure what you meant by "dynamically change the end-time of a window. If you are referring to dynamically determines the firing time of the window, then it fits into the description of session window [1]: 
If you want to handle window end time dynamically, one way of which I can think of is the dynamic gap, session window [1] approach. with which you can specify the end-time of a window based on input elements. Provided that you are maintaining a session window. 
Another way to look at it is through the Flink-CEP library [2]. 

Thanks,
Rong



On Tue, Apr 23, 2019 at 8:19 AM M Singh <[hidden email]> wrote:
Hi:

I am working on a project and need to change the end time of the window dynamically.  I want to find out if the end time of the window is used internally (for sorting windows/etc) except for handling watermarks that would cause problems if the end time was changed during run time after the window has been created even if no new event has arrived for that window.

I don't want to use GlobalWindow since from my understanding it never gets destroyed.

If there is any alternate way of dealing with this, please let me know.

Thanks

Mans
M Singh
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Apache Flink - Question about dynamically changing window end time at run time

M Singh
Thanks Sameer/Rong:

As Fabian and you have mentioned, the window still sticks around forever for global window, so I am trying avoid that scenario.

Fabian & Flink team - do you have any insights into what would happen if I create a window and the later change it's end time during the stream processing ?  Would it mess up any internal state/processing that uses the end time when the window was first created ?  If there is any other consideration to keep in mind, please let me know.

Thanks again.

On Wednesday, April 24, 2019, 1:29:18 PM EDT, Rong Rong <[hidden email]> wrote:


Hi Mans,

Sameer is correct. if you would like to control window triggering based on other elements that does not belong to this window (in a keyed stream context) then this is probably the best way to approach. 

I think you've also posted in another thread that describes what will be left after fire-and-purge [1]. As Fabian stated: the only thing that might've left after is the window (which is the 2 long values indicate the start/end) and the trigger object. But you are right it might eventually filled up memory.

Another approach is to implement your own operator that handles all these internally by your user code. This would require you to replicate many of the window operator logic though.

Thanks,
Rong


On Wed, Apr 24, 2019 at 5:02 AM Sameer W <[hidden email]> wrote:
Global Windows is fine for this use case. I have used the same strategy. You just define custom evictors and triggers and you are all good. Windows are managed by keys, so as such as long as events are evicted from the window, that counts towards reclaiming memory for the key+window combination. Plus there is just window per key with Global Windows. 

On Wed, Apr 24, 2019 at 7:47 AM M Singh <[hidden email]> wrote:
Hi Rong:

Thanks for your answer.

From what I understand the dynamic gap session windows are also created when the event is encountered.  I need to be able to change the window end time at a later time based on what other events are in that window.  One way to do this is to use GlobalWindows but then these are never deleted.

Regarding CEP option - I believe that CEP patterns cannot be changed dynamically once they've been complied which limits it usage.

Please feel free to correct me.

Thanks for your help and pointers.

On Tuesday, April 23, 2019, 8:12:56 PM EDT, Rong Rong <[hidden email]> wrote:


Hi Mans,

I am not sure what you meant by "dynamically change the end-time of a window. If you are referring to dynamically determines the firing time of the window, then it fits into the description of session window [1]: 
If you want to handle window end time dynamically, one way of which I can think of is the dynamic gap, session window [1] approach. with which you can specify the end-time of a window based on input elements. Provided that you are maintaining a session window. 
Another way to look at it is through the Flink-CEP library [2]. 

Thanks,
Rong



On Tue, Apr 23, 2019 at 8:19 AM M Singh <[hidden email]> wrote:
Hi:

I am working on a project and need to change the end time of the window dynamically.  I want to find out if the end time of the window is used internally (for sorting windows/etc) except for handling watermarks that would cause problems if the end time was changed during run time after the window has been created even if no new event has arrived for that window.

I don't want to use GlobalWindow since from my understanding it never gets destroyed.

If there is any alternate way of dealing with this, please let me know.

Thanks

Mans
Fabian Hueske-2
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Apache Flink - Question about dynamically changing window end time at run time

Fabian Hueske-2
Hi Mans,

I don't know if that would work or not. Would need to dig into the source code for that.

TBH, I would recommend to check if you can implement the logic using a (Keyed-)ProcessFunction.
IMO, process functions are a lot easier to reason about than Flink's windowing framework.
You can manage state and timer all by yourself and make sure everything is properly cleaned up.

Best,
Fabian


Am So., 28. Apr. 2019 um 16:31 Uhr schrieb M Singh <[hidden email]>:
Thanks Sameer/Rong:

As Fabian and you have mentioned, the window still sticks around forever for global window, so I am trying avoid that scenario.

Fabian & Flink team - do you have any insights into what would happen if I create a window and the later change it's end time during the stream processing ?  Would it mess up any internal state/processing that uses the end time when the window was first created ?  If there is any other consideration to keep in mind, please let me know.

Thanks again.

On Wednesday, April 24, 2019, 1:29:18 PM EDT, Rong Rong <[hidden email]> wrote:


Hi Mans,

Sameer is correct. if you would like to control window triggering based on other elements that does not belong to this window (in a keyed stream context) then this is probably the best way to approach. 

I think you've also posted in another thread that describes what will be left after fire-and-purge [1]. As Fabian stated: the only thing that might've left after is the window (which is the 2 long values indicate the start/end) and the trigger object. But you are right it might eventually filled up memory.

Another approach is to implement your own operator that handles all these internally by your user code. This would require you to replicate many of the window operator logic though.

Thanks,
Rong


On Wed, Apr 24, 2019 at 5:02 AM Sameer W <[hidden email]> wrote:
Global Windows is fine for this use case. I have used the same strategy. You just define custom evictors and triggers and you are all good. Windows are managed by keys, so as such as long as events are evicted from the window, that counts towards reclaiming memory for the key+window combination. Plus there is just window per key with Global Windows. 

On Wed, Apr 24, 2019 at 7:47 AM M Singh <[hidden email]> wrote:
Hi Rong:

Thanks for your answer.

From what I understand the dynamic gap session windows are also created when the event is encountered.  I need to be able to change the window end time at a later time based on what other events are in that window.  One way to do this is to use GlobalWindows but then these are never deleted.

Regarding CEP option - I believe that CEP patterns cannot be changed dynamically once they've been complied which limits it usage.

Please feel free to correct me.

Thanks for your help and pointers.

On Tuesday, April 23, 2019, 8:12:56 PM EDT, Rong Rong <[hidden email]> wrote:


Hi Mans,

I am not sure what you meant by "dynamically change the end-time of a window. If you are referring to dynamically determines the firing time of the window, then it fits into the description of session window [1]: 
If you want to handle window end time dynamically, one way of which I can think of is the dynamic gap, session window [1] approach. with which you can specify the end-time of a window based on input elements. Provided that you are maintaining a session window. 
Another way to look at it is through the Flink-CEP library [2]. 

Thanks,
Rong



On Tue, Apr 23, 2019 at 8:19 AM M Singh <[hidden email]> wrote:
Hi:

I am working on a project and need to change the end time of the window dynamically.  I want to find out if the end time of the window is used internally (for sorting windows/etc) except for handling watermarks that would cause problems if the end time was changed during run time after the window has been created even if no new event has arrived for that window.

I don't want to use GlobalWindow since from my understanding it never gets destroyed.

If there is any alternate way of dealing with this, please let me know.

Thanks

Mans
M Singh
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Apache Flink - Question about dynamically changing window end time at run time

M Singh
Sounds great Fabian. 

I was just trying to see if I can use higher level datastream apis. 

I appreciate your advice and help. 

Mans

On Monday, April 29, 2019, 5:41:36 AM EDT, Fabian Hueske <[hidden email]> wrote:


Hi Mans,

I don't know if that would work or not. Would need to dig into the source code for that.

TBH, I would recommend to check if you can implement the logic using a (Keyed-)ProcessFunction.
IMO, process functions are a lot easier to reason about than Flink's windowing framework.
You can manage state and timer all by yourself and make sure everything is properly cleaned up.

Best,
Fabian


Am So., 28. Apr. 2019 um 16:31 Uhr schrieb M Singh <[hidden email]>:
Thanks Sameer/Rong:

As Fabian and you have mentioned, the window still sticks around forever for global window, so I am trying avoid that scenario.

Fabian & Flink team - do you have any insights into what would happen if I create a window and the later change it's end time during the stream processing ?  Would it mess up any internal state/processing that uses the end time when the window was first created ?  If there is any other consideration to keep in mind, please let me know.

Thanks again.

On Wednesday, April 24, 2019, 1:29:18 PM EDT, Rong Rong <[hidden email]> wrote:


Hi Mans,

Sameer is correct. if you would like to control window triggering based on other elements that does not belong to this window (in a keyed stream context) then this is probably the best way to approach. 

I think you've also posted in another thread that describes what will be left after fire-and-purge [1]. As Fabian stated: the only thing that might've left after is the window (which is the 2 long values indicate the start/end) and the trigger object. But you are right it might eventually filled up memory.

Another approach is to implement your own operator that handles all these internally by your user code. This would require you to replicate many of the window operator logic though.

Thanks,
Rong


On Wed, Apr 24, 2019 at 5:02 AM Sameer W <[hidden email]> wrote:
Global Windows is fine for this use case. I have used the same strategy. You just define custom evictors and triggers and you are all good. Windows are managed by keys, so as such as long as events are evicted from the window, that counts towards reclaiming memory for the key+window combination. Plus there is just window per key with Global Windows. 

On Wed, Apr 24, 2019 at 7:47 AM M Singh <[hidden email]> wrote:
Hi Rong:

Thanks for your answer.

From what I understand the dynamic gap session windows are also created when the event is encountered.  I need to be able to change the window end time at a later time based on what other events are in that window.  One way to do this is to use GlobalWindows but then these are never deleted.

Regarding CEP option - I believe that CEP patterns cannot be changed dynamically once they've been complied which limits it usage.

Please feel free to correct me.

Thanks for your help and pointers.

On Tuesday, April 23, 2019, 8:12:56 PM EDT, Rong Rong <[hidden email]> wrote:


Hi Mans,

I am not sure what you meant by "dynamically change the end-time of a window. If you are referring to dynamically determines the firing time of the window, then it fits into the description of session window [1]: 
If you want to handle window end time dynamically, one way of which I can think of is the dynamic gap, session window [1] approach. with which you can specify the end-time of a window based on input elements. Provided that you are maintaining a session window. 
Another way to look at it is through the Flink-CEP library [2]. 

Thanks,
Rong



On Tue, Apr 23, 2019 at 8:19 AM M Singh <[hidden email]> wrote:
Hi:

I am working on a project and need to change the end time of the window dynamically.  I want to find out if the end time of the window is used internally (for sorting windows/etc) except for handling watermarks that would cause problems if the end time was changed during run time after the window has been created even if no new event has arrived for that window.

I don't want to use GlobalWindow since from my understanding it never gets destroyed.

If there is any alternate way of dealing with this, please let me know.

Thanks

Mans
Free forum by Nabble Edit this page