Github user WeichenXu123 commented on the issue: https://github.com/apache/spark/pull/19661 So why do you include the class such as `org.apache.spark.ml.feature.Instance`. You can look into a lot of algos, in `ml` package (not `mllib`), still use something like `RDD[Instance]` and cache them, (these algos only interface use dataframe). such as LinearRegression, LogisticRegression and so on.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org