gaoyunhaii commented on pull request #11: URL: https://github.com/apache/flink-ml/pull/11#issuecomment-934234370
Very thanks @becketqin for the review! > I personally feel it is a little confusing to have the concept of draft subgraph. Would the following work? It seems to me this would not work since the `StreamExecutionEnvironment` is created by the users, like ``` StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment(); // add some transformations to the env // add the iteration ``` Unless we make `IterationExecutionEnvironment` visible to the users and ask users to create the environment initially, we might not be able to achieve both between `switch to a second environment` and `add the transformation back to the original environment`? The only possible method to me might be have a second `IterationExecutionEnvironment`, when adding transformation to it, it adds the translated transformation into the original environment, but it seems to me have too much difference with the current `DraftExecutionEnvironment`? Since not all transformation would be added by the `addOperator`, the new `IterationExecutionEnvironment` would still need to search for a subgraph that need to be translated. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
