[jira] [Created] (FLINK-2497) Rework streaming iteration flow control

2015-08-07 Thread Gyula Fora (JIRA)
Gyula Fora created FLINK-2497: - Summary: Rework streaming iteration flow control Key: FLINK-2497 URL: https://issues.apache.org/jira/browse/FLINK-2497 Project: Flink Issue Type: Improvement

Re: streaming iteration

2015-07-27 Thread Gyula Fóra
Hey, The JobGraph that is executed cannot have cycles (due to scheduling reasons), that's why there is no edge between the head and tail operators. What we do instead is we add an extra source to the head and a sink to the tail (with the same parallelism), and the feedback data is passed outside o

streaming iteration

2015-07-27 Thread MaGuoWei
Hi,guysI am reading the code about streaming iteration.I find there is no StreamEdge the StreamIterationHead and StreamIterationTail.Why? Because the graph can't have any ring?Or any other concerns? thanks

Re: Rework of streaming iteration API

2015-07-07 Thread Paris Carbone
;>>>> creation. For example: >>>>> >>>>> DataStream source = .. >>>>> DataStream mapper = source.map(noOpMapper) >>>>> DataStream feedback = mapper.filter(...) >>>>> source.union(feedback) >>>>

Re: Rework of streaming iteration API

2015-07-07 Thread Kostas Tzoumas
back = mapper.filter(...) > > > > source.union(feedback) > > > > > > > > This would basically mean that a DataStream is mutable and can be > > > extended > > > > after creation with more streams. > > > > > > > > On Tue, 7 Jul 2015

Re: Rework of streaming iteration API

2015-07-07 Thread Gyula Fóra
ataStream is mutable and can be > > extended > > > after creation with more streams. > > > > > > On Tue, 7 Jul 2015 at 16:12 Aljoscha Krettek > > wrote: > > > > > > > I think this would be good yes. I was just about to open an Issue

Re: Rework of streaming iteration API

2015-07-07 Thread Stephan Ewen
at 16:12 Aljoscha Krettek > wrote: > > > > > I think this would be good yes. I was just about to open an Issue for > > > changing the Streaming Iteration API. :D > > > > > > Then we should also make the implementation very straightforward and >

Re: Rework of streaming iteration API

2015-07-07 Thread Gyula Fóra
hink this would be good yes. I was just about to open an Issue for > > changing the Streaming Iteration API. :D > > > > Then we should also make the implementation very straightforward and > > simple, right now, the implementation of the iterations is all over the > >

Re: Rework of streaming iteration API

2015-07-07 Thread Aljoscha Krettek
after creation with more streams. On Tue, 7 Jul 2015 at 16:12 Aljoscha Krettek wrote: > I think this would be good yes. I was just about to open an Issue for > changing the Streaming Iteration API. :D > > Then we should also make the implementation very straightforward and > si

Re: Rework of streaming iteration API

2015-07-07 Thread Gyula Fóra
; >> I think this would be good yes. I was just about to open an Issue for >> changing the Streaming Iteration API. :D >> >> Then we should also make the implementation very straightforward and >> simple, right now, the implementation of the iterations is all ov

Re: Rework of streaming iteration API

2015-07-07 Thread Gyula Fóra
ink this would be good yes. I was just about to open an Issue for > changing the Streaming Iteration API. :D > > Then we should also make the implementation very straightforward and > simple, right now, the implementation of the iterations is all over the > place. > > On Tue, 7

Re: Rework of streaming iteration API

2015-07-07 Thread Kostas Tzoumas
+1 for rethinking the iterations in DataStream However, wouldn't this proposal allow the definition of arbitrary loops (e.g., nested loops) that are not well behaved afaik? On Tue, Jul 7, 2015 at 4:12 PM, Stephan Ewen wrote: > I see that the newly proposed API makes some things easier to define

Re: Rework of streaming iteration API

2015-07-07 Thread Aljoscha Krettek
I think this would be good yes. I was just about to open an Issue for changing the Streaming Iteration API. :D Then we should also make the implementation very straightforward and simple, right now, the implementation of the iterations is all over the place. On Tue, 7 Jul 2015 at 15:57 Gyula

Re: Rework of streaming iteration API

2015-07-07 Thread Stephan Ewen
I see that the newly proposed API makes some things easier to define. There is one source of confusion, though, in my opinion: The new API suggests that the data stream actually refers to the operator that created it. The "DataStream mapper = source.map(noOpMapper)" here refers to the map operato

Rework of streaming iteration API

2015-07-07 Thread Gyula Fóra
Hey, Along with the suggested changes to the streaming API structure I think we should also rework the "iteration" api. Currently the iteration api tries to mimic the syntax of the batch API while the runtime behaviour is quite different. What we create instead of iterations is really just cyclic

[jira] [Created] (FLINK-1854) Planvisualizer cannot visualize the streaming iteration example

2015-04-09 Thread Gyula Fora (JIRA)
Gyula Fora created FLINK-1854: - Summary: Planvisualizer cannot visualize the streaming iteration example Key: FLINK-1854 URL: https://issues.apache.org/jira/browse/FLINK-1854 Project: Flink

[jira] [Created] (FLINK-1686) Streaming iteration heads cannot be instantiated

2015-03-11 Thread Paris Carbone (JIRA)
Paris Carbone created FLINK-1686: Summary: Streaming iteration heads cannot be instantiated Key: FLINK-1686 URL: https://issues.apache.org/jira/browse/FLINK-1686 Project: Flink Issue Type