You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@flink.apache.org by Abhijeet Kumar <ab...@sentienz.com> on 2018/11/26 12:13:04 UTC

Flink join stream where one stream is coming 5 minutes late

Hello Team,

I've to join two stream where one stream is coming late. So, I planned doing it by creating two windows, for first window the size will be 5 minutes and for the other it will be 10 minutes. Then, I'll perform the join operation. Is my solution correct because I don't know whether one window will sync with other or not because one is getting finished earlier and another is late. If there is a better way to achieve the same, please let me know. 

Thanks,

	
Abhijeet Kumar
Software Development Engineer,
Sentienz Solutions Pvt Ltd
Cognitive Data Platform - Perceive the Data !
abhijeet.kumar@sentienz.com <ma...@sentienz.com> |www.sentienz.com <http://www.sentienz.com/> | Bengaluru



Re: Flink join stream where one stream is coming 5 minutes late

Posted by Abhijeet Kumar <ab...@sentienz.com>.
Thanks for the response, Xingcan. I got your point and the whole idea.

> On 27-Nov-2018, at 12:15 AM, Xingcan Cui <xi...@gmail.com> wrote:
> 
> Hi Abhijeet,
> 
> If you want to perform window-join in the DataStream API, the window configurations on both sides must be exactly the same.
> 
> For your case, maybe you can try adding a 5 mins delay on event times (and watermarks) of the faster stream. 
> 
> Hope that helps.
> 
> Best,
> Xingcan
> 
> 
>> On Nov 26, 2018, at 12:44 PM, Abhijeet Kumar <abhijeet.kumar@sentienz.com <ma...@sentienz.com>> wrote:
>> 
>> I already mentioned above, one is coming late by 5 mins. Maybe my approach isn't correct, and I asked to correct me if I'm wrong. 
>> 
>>> On 26-Nov-2018, at 5:56 PM, Taher Koitawala <taher.koitawala@gslab.com <ma...@gslab.com>> wrote:
>>> 
>>> May I ask why you want to have 2 differences between window time? What's the use case?
>>> 
>>> On Mon 26 Nov, 2018, 5:53 PM Abhijeet Kumar <abhijeet.kumar@sentienz.com <ma...@sentienz.com> wrote:
>>> Hello Team,
>>> 
>>> I've to join two stream where one stream is coming late. So, I planned doing it by creating two windows, for first window the size will be 5 minutes and for the other it will be 10 minutes. Then, I'll perform the join operation. Is my solution correct because I don't know whether one window will sync with other or not because one is getting finished earlier and another is late. If there is a better way to achieve the same, please let me know. 
>>> 
>>> Thanks,
>>> 
>>> 	
>>> Abhijeet Kumar
>>> Software Development Engineer,
>>> Sentienz Solutions Pvt Ltd
>>> Cognitive Data Platform - Perceive the Data !
>>> abhijeet.kumar@sentienz.com <ma...@sentienz.com> |www.sentienz.com <http://www.sentienz.com/> | Bengaluru
>>> 
>>> 
>> 
> 


Re: Flink join stream where one stream is coming 5 minutes late

Posted by Xingcan Cui <xi...@gmail.com>.
Hi Abhijeet,

If you want to perform window-join in the DataStream API, the window configurations on both sides must be exactly the same.

For your case, maybe you can try adding a 5 mins delay on event times (and watermarks) of the faster stream. 

Hope that helps.

Best,
Xingcan


> On Nov 26, 2018, at 12:44 PM, Abhijeet Kumar <ab...@sentienz.com> wrote:
> 
> I already mentioned above, one is coming late by 5 mins. Maybe my approach isn't correct, and I asked to correct me if I'm wrong. 
> 
>> On 26-Nov-2018, at 5:56 PM, Taher Koitawala <taher.koitawala@gslab.com <ma...@gslab.com>> wrote:
>> 
>> May I ask why you want to have 2 differences between window time? What's the use case?
>> 
>> On Mon 26 Nov, 2018, 5:53 PM Abhijeet Kumar <abhijeet.kumar@sentienz.com <ma...@sentienz.com> wrote:
>> Hello Team,
>> 
>> I've to join two stream where one stream is coming late. So, I planned doing it by creating two windows, for first window the size will be 5 minutes and for the other it will be 10 minutes. Then, I'll perform the join operation. Is my solution correct because I don't know whether one window will sync with other or not because one is getting finished earlier and another is late. If there is a better way to achieve the same, please let me know. 
>> 
>> Thanks,
>> 
>> 	
>> Abhijeet Kumar
>> Software Development Engineer,
>> Sentienz Solutions Pvt Ltd
>> Cognitive Data Platform - Perceive the Data !
>> abhijeet.kumar@sentienz.com <ma...@sentienz.com> |www.sentienz.com <http://www.sentienz.com/> | Bengaluru
>> 
>> 
> 


Re: Flink join stream where one stream is coming 5 minutes late

Posted by Abhijeet Kumar <ab...@sentienz.com>.
I already mentioned above, one is coming late by 5 mins. Maybe my approach isn't correct, and I asked to correct me if I'm wrong. 

> On 26-Nov-2018, at 5:56 PM, Taher Koitawala <ta...@gslab.com> wrote:
> 
> May I ask why you want to have 2 differences between window time? What's the use case?
> 
> On Mon 26 Nov, 2018, 5:53 PM Abhijeet Kumar <abhijeet.kumar@sentienz.com <ma...@sentienz.com> wrote:
> Hello Team,
> 
> I've to join two stream where one stream is coming late. So, I planned doing it by creating two windows, for first window the size will be 5 minutes and for the other it will be 10 minutes. Then, I'll perform the join operation. Is my solution correct because I don't know whether one window will sync with other or not because one is getting finished earlier and another is late. If there is a better way to achieve the same, please let me know. 
> 
> Thanks,
> 
> 	
> Abhijeet Kumar
> Software Development Engineer,
> Sentienz Solutions Pvt Ltd
> Cognitive Data Platform - Perceive the Data !
> abhijeet.kumar@sentienz.com <ma...@sentienz.com> |www.sentienz.com <http://www.sentienz.com/> | Bengaluru
> 
> 


Re: Flink join stream where one stream is coming 5 minutes late

Posted by Taher Koitawala <ta...@gslab.com>.
May I ask why you want to have 2 differences between window time? What's
the use case?

On Mon 26 Nov, 2018, 5:53 PM Abhijeet Kumar <abhijeet.kumar@sentienz.com
wrote:

> Hello Team,
>
> I've to join two stream where one stream is coming late. So, I planned
> doing it by creating two windows, for first window the size will be 5
> minutes and for the other it will be 10 minutes. Then, I'll perform the
> join operation. Is my solution correct because I don't know whether one
> window will sync with other or not because one is getting finished earlier
> and another is late. If there is a better way to achieve the same, please
> let me know.
>
> Thanks,
>
>
> *Abhijeet Kumar*
> Software Development Engineer,
> Sentienz Solutions Pvt Ltd
> Cognitive Data Platform - Perceive the Data !
> abhijeet.kumar@sentienz.com |www.sentienz.com | Bengaluru
>
>
>