[GitHub] [flink] xuyang1706 commented on a change in pull request #9355: [FLINK-13577][ml] Add an util class to build result row and generate …

[GitBox <gi...@apache.org>]

xuyang1706 commented on a change in pull request #9355: [FLINK-13577][ml] Add an util class to build result row and generate …
URL: https://github.com/apache/flink/pull/9355#discussion_r353542081
 
 

 ##########
 File path: flink-ml-parent/flink-ml-lib/src/main/java/org/apache/flink/ml/common/utils/OutputColsHelper.java
 ##########
 @@ -0,0 +1,186 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+package org.apache.flink.ml.common.utils;
+
+import org.apache.flink.api.common.typeinfo.TypeInformation;
+import org.apache.flink.table.api.TableSchema;
+import org.apache.flink.types.Row;
+
+import org.apache.commons.lang3.ArrayUtils;
+
+import java.io.Serializable;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.HashSet;
+
+/**
+ * Util for generating output schema when doing prediction or transformation.
+ *
+ * <p>Input:
+ * 1) Schema of input data being predicted or transformed.
+ * 2) Output column names of the prediction/transformation operator.
+ * 3) Output column types of the prediction/transformation operator.
+ * 4) Reserved column names, which is a subset of input data's column names that we want to preserve.
+ *
+ * <p>Output:
+ * 1）The result data schema. The result data is a combination of the preserved columns and the operator's
+ * output columns.
+ *
+ * <p>Several rules are followed:
+ * 1) If reserved columns are not given, then all columns of input data is reserved.
+ * 2）The reserved columns are arranged ahead of the operator's output columns in the final output.
+ * 3) If some of the reserved column names overlap with those of operator's output columns, then the operator's
 
 Review comment:
   Thanks for your advice. As far as we know from our practice, in the domain of ML, reserving input columns by default could be a better design. This is more than the framework itself, but also about user experience. For example, data preprocessing, always the first step to ML, often performs transformation on multiple columns one by one. So there may be PilelineStage_1 for feature scaling on column_1, PilelineStage_2 for binarizing on column_2 and more. This would be redundant to force user to add a flag for each of these steps.

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services