Re: New DoFn and WindowedValue/WinowingInternals

2016-12-11 Thread Eugene Kirpichov
Hi Amit, Yes, this is correct. Part of the motivation for this is that DoFn API is user-facing, and the compressed representation of windowed elements (e.g. access to all windows of an element), as well as the ability to emit directly into a specified window, is an implementation detail of the

Re: New DoFn and WindowedValue/WinowingInternals

2016-12-11 Thread Kenneth Knowles
You've got it right. My recommendations is to just directly implement it for the Spark runner. It will often actually clean things up a bit. Here's the analogous change for the Flink runner: https://github.com/apache/incubator-beam/pull/1435/files. With GABW, I tried going through the process of

Re: New DoFn and WindowedValue/WinowingInternals

2016-12-11 Thread Amit Sela
So basically using new DoFn with *SplittableParDo*, *Window.Bound*, and *GroupAlsoByWindow* requires a custom implementation by per runner as they are not handled by DoFn anymore, right ? On Sun, Dec 11, 2016 at 3:42 PM Eugene Kirpichov wrote: > Hi Amit, I'll

Re: New DoFn and WindowedValue/WinowingInternals

2016-12-11 Thread Eugene Kirpichov
Hi Amit, I'll comment in more detail later, but meanwhile please take a look at https://github.com/apache/incubator-beam/pull/1565 There is a small amount of relevant changes to spark runner. Take a look at implementation of SplittableParDo (already committed) in particular ProcessFn and it's