Re: [QUESTION] Should DoFns be able to get the watermark?

2019-04-09 Thread Lukasz Cwik
Yes, those are the ones I was referring to. On Tue, Apr 9, 2019 at 4:08 PM Pablo Estrada wrote: > Yup! : ) - I think > > On Tue, Apr 9, 2019 at 3:52 PM Brian Hulette wrote: > >> Are these the blog posts? >> >> https://beam.apache.org/blog/2017/02/13/stateful-processing.html >>

Re: [QUESTION] Should DoFns be able to get the watermark?

2019-04-09 Thread Pablo Estrada
Yup! : ) - I think On Tue, Apr 9, 2019 at 3:52 PM Brian Hulette wrote: > Are these the blog posts? > > https://beam.apache.org/blog/2017/02/13/stateful-processing.html > https://beam.apache.org/blog/2017/08/28/timely-processing.html > > On Tue, Apr 9, 2019 at 3:41 PM Pablo Estrada wrote: > >>

Re: [QUESTION] Should DoFns be able to get the watermark?

2019-04-09 Thread Brian Hulette
Are these the blog posts? https://beam.apache.org/blog/2017/02/13/stateful-processing.html https://beam.apache.org/blog/2017/08/28/timely-processing.html On Tue, Apr 9, 2019 at 3:41 PM Pablo Estrada wrote: > sunds good. Thanks guys <3 > > On Tue, Apr 9, 2019 at 3:19 PM Lukasz Cwik wrote:

Re: [QUESTION] Should DoFns be able to get the watermark?

2019-04-09 Thread Pablo Estrada
sunds good. Thanks guys <3 On Tue, Apr 9, 2019 at 3:19 PM Lukasz Cwik wrote: > UnboundedSources and SplittableDoFns report watermarks which the runner > uses to compute how much the watermark could advance if it processed some > outstanding work. But it is always upto the runner to choose

Re: [QUESTION] Should DoFns be able to get the watermark?

2019-04-09 Thread Lukasz Cwik
UnboundedSources and SplittableDoFns report watermarks which the runner uses to compute how much the watermark could advance if it processed some outstanding work. But it is always upto the runner to choose when the watermark advances. The runner could process each work item in watermark priority

Re: [QUESTION] Should DoFns be able to get the watermark?

2019-04-09 Thread Kenneth Knowles
In state & timers and new DoFn in the past It was an explicit decision to not allow direct observation of the watermark, but only to set a timer in event time. Is there a design doc I can read to catch up? Kenn On Tue, Apr 9, 2019 at 1:44 PM Lukasz Cwik wrote: > WatermarkReporterParam is about

Re: [QUESTION] Should DoFns be able to get the watermark?

2019-04-09 Thread Pablo Estrada
hi Luke, thanks for the prompt reply: ) That makes sense. I think I'll go back to my cave to read a bunch about streaming. : ) I was looking for this to try to write a sequence generator for Python in streaming, and I was trying to debug what was going on. I was trying to allow the DoFn to

Re: [QUESTION] Should DoFns be able to get the watermark?

2019-04-09 Thread Lukasz Cwik
WatermarkReporterParam is about reporting the watermark. The main usecase is for SplittableDoFns to be able to report the data watermark. The watermark is per input and output of a DoFn. Also each bundle being processed has its local watermarks while the runner computes the global watermark. The