c4emmmm commented on issue #8402: [FLINK-12473][ml] Add the interface of ML pipeline and ML lib URL: https://github.com/apache/flink/pull/8402#issuecomment-494672243 As mentioned in the mail "[DISCUSS] FLIP-39: Flink ML pipeline and ML libs" in flink-dev, we have to add a TableEnvironment as input argument in Estimator.fit() and Transformer.transform(). This is because it's necessary to register a function, and may use to judge whether the pipeline is running in a batch or streaming environment since some algorithm only support batch training. While it is impossible to get the TableEnvironment via the only input Table without a flink-table-planner dependency, which should not be touched by users, so we have to add this as input. Another way is to get the TableEnvironment via a static method, but it is not safe since users may create a TableEnvironment without using this method, and cause unknown problems. Adding the input argument should not change much the experience of users and library developers. Please let me know if anyone has questions about this.
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
