Github user jkbradley commented on the pull request: https://github.com/apache/spark/pull/4460#issuecomment-74609769 I like the current sketch but also want to think about it more. A few thoughts: I'm not quite clear on how the Array of Attributes in FeatureAttributes corresponds to the columns of the DataFrame. Is it one-to-one, or will Attributes be nested? (I'm basically thinking about groups of features, especially individual features grouped into vectors.) How will propagation of feature names work? Will we try to impose a standard, such as Transformers maintaining the same (or a modified) feature name whenever possible? By the way, do we want to call this "FeatureAttributes," or should we name it something like "ColumnAttributes" so it more obviously applies to other types of columns like labels, users, products, etc.? +1 for moving FeatureType from mllib.tree to attribute. It should be more general.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org