weibozhao commented on code in PR #156: URL: https://github.com/apache/flink-ml/pull/156#discussion_r992921852
########## flink-ml-lib/src/main/java/org/apache/flink/ml/feature/vectorassembler/VectorAssembler.java: ########## @@ -47,10 +47,15 @@ /** * A Transformer which combines a given list of input columns into a vector column. Types of input - * columns must be either vector or numerical value. + * columns must be either vector or numerical types. The elements assembled in the same column must + * have the same size. The operator deals with null values or records with wrong sizes according to + * the strategy specified by the {@link HasHandleInvalid} parameter as follows: * - * <p>The `keep` option of {@link HasHandleInvalid} means that we output bad rows with output column - * set to null. + * <p>The `keep` option means that we do the assembling action without checking the vector size. Review Comment: assembling [1,2], [3,4,5] with sizes 2,2 may get [1, 2, 3, 4] (trim to fit size) assembling [1,2], [3,4,5] with sizes 2,2 may get [1, 2, NaN, NaN] assembling [1,2], null with sizes 2,2 may get [1, 2, NaN, NaN] assembling [1,2], null with sizes 2,2 may get [1, 2, 0, 0] -- 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: issues-unsubscr...@flink.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org