jiangxin369 commented on code in PR #162: URL: https://github.com/apache/flink-ml/pull/162#discussion_r1007692506
########## flink-ml-lib/src/main/java/org/apache/flink/ml/common/util/QuantileSummary.java: ########## @@ -0,0 +1,400 @@ +/* + * 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.util; + +import org.apache.flink.annotation.Internal; +import org.apache.flink.util.Preconditions; + +import java.io.Serializable; +import java.util.ArrayList; +import java.util.Arrays; +import java.util.Collections; +import java.util.HashMap; +import java.util.LinkedList; +import java.util.List; +import java.util.Map; +import java.util.stream.Collectors; +import java.util.stream.IntStream; + +/** + * Helper class to compute an approximate quantile summary. This implementation is based on the + * algorithm proposed in the paper: "Space-efficient Online Computation of Quantile Summaries" by + * Greenwald, Michael and Khanna, Sanjeev. (https://doi.org/10.1145/375663.375670) + */ +@Internal +public class QuantileSummary implements Serializable { + + /** The target relative error. */ + private final double relativeError; + + /** + * The compression threshold. After the internal buffer of statistics crosses this size, it + * attempts to compress the statistics together. + */ + private final int compressThreshold; + + /** The count of all the elements inserted to be calculated. */ + private final long count; + + /** A buffer of quantile statistics. */ + private final List<StatsTuple> sampled; + + /** The default size of head buffer. */ + private static final int DEFAULT_HEAD_SIZE = 50000; + + /** The default compression threshold. */ + private static final int DEFAULT_COMPRESS_THRESHOLD = 10000; + + /** A buffer of the latest samples seen so far. */ + private List<Double> headBuffer = new ArrayList<>(DEFAULT_HEAD_SIZE); Review Comment: Indeed `double[]` is more memory efficient, but after repeated thinking I think leave it as an empty ArrayList is better, it is more friendly to small dataset, which is the same to Spark. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
