[jira] [Commented] (BEAM-6372) Direct Runner should marshal data in a similar way to Dataflow runner

["Robert Burke (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/BEAM-6372?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16736201#comment-16736201 ] 

Robert Burke commented on BEAM-6372:
------------------------------------

Off the top of my head, I think I agree with the assessment, the error message would ideally be very clear about what's missing and possibly why. One reason it hasn't been done this way is that for testing  with ptest and `go test`, it's inconvenient to have the registration scaffolding everywhere. In particular we'd probably want to maintain that ease, but still have the semantics check everything correctly, likely with a --test flag that's used by default in the testing harness.

> Direct Runner should marshal data in a similar way to Dataflow runner
> ---------------------------------------------------------------------
>
>                 Key: BEAM-6372
>                 URL: https://issues.apache.org/jira/browse/BEAM-6372
>             Project: Beam
>          Issue Type: Improvement
>          Components: runner-direct, sdk-go
>            Reporter: Andrew Brampton
>            Assignee: Robert Burke
>            Priority: Major
>
> I would test my pipeline using the direct runner, and it would happily run on a sample. When I ran it on the Dataflow runner, it'll run for a hour, then get to a stage that would crash like so:
>  
> {quote}java.util.concurrent.ExecutionException: java.lang.RuntimeException: Error received from SDK harness for instruction -224: execute failed: panic: reflect: Call using main.HistogramResult as type struct \{ Key string "json:\"key\""; Files []string "json:\"files\""; Histogram palette.ColorHistogram "json:\"histogram,omitempty\""; Palette []struct { R uint8; G uint8; B uint8; A uint8 } "json:\"palette\"" } goroutine 70 [running]:{quote}
> This was because I forgot to register my HistogramResult type.
> It would be useful if the direct runner tried to marshal and unmarshal all types, to help expose issues like this earlier.
> Also, when running on Dataflow, the value of flags, and captured variables, would be the empty/default value. It would be good if direct also caused this behaviour. For example:
> {code}
> prefix := “X”
> s = s.Scope(“Prefix ” + prefix)
> c = beam.ParDo(s, func(value string) string {
> 	return prefix + value
> }, c)
> {code}
> Will work prefix "X" on the Direct runner, but will prefix "" on Dataflow. Subtle behaviour, but I suspect the direct runner could expose this if it marshalled and unmarshalled the func like the dataflow runner.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)