[ 
https://issues.apache.org/jira/browse/BEAM-1982?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16000627#comment-16000627
 ] 

Borisa Zivkovic edited comment on BEAM-1982 at 5/8/17 11:43 AM:
----------------------------------------------------------------

Do you want this to be done 1) only at the level of 
org.apache.beam.runners.core.construction.SingleInputOutputOverrideFactory<InputT,
 OutputT, TransformT> and its subclasses by adding common code to 
SingleInputOutputOverrideFactory or 

2) we should introduce new abstract class implementing 
org.apache.beam.sdk.runners.PTransformOverrideFactory<InputT, OutputT, 
TransformT> and adding this common code?

Second option is more intrusive since it requires changing some classe to 
extend abstract class instead of implementing interface like they do today (and 
this in some cases might be impossible).


was (Author: zborisha):
Do you want this to be done 1) only at the level of 
org.apache.beam.runners.core.construction.SingleInputOutputOverrideFactory<InputT,
 OutputT, TransformT> and its subclasses) or 2) we should introduce new 
abstract class implementing 
org.apache.beam.sdk.runners.PTransformOverrideFactory<InputT, OutputT, 
TransformT> and adding this common code?

Two is more intrusive.

> Refactor PTransformOverrideFactories to remove code duplication
> ---------------------------------------------------------------
>
>                 Key: BEAM-1982
>                 URL: https://issues.apache.org/jira/browse/BEAM-1982
>             Project: Beam
>          Issue Type: Bug
>          Components: runner-core
>            Reporter: Thomas Groh
>            Priority: Minor
>              Labels: newbie, starter
>
> The majority of the factories take in a single PCollection as input and 
> produce a single PCollection as output; the code to extract the single input 
> can be invisible as part of the class hierarchy, rather than explicitly 
> called in every override.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Reply via email to