Aitozi commented on code in PR #21257: URL: https://github.com/apache/flink/pull/21257#discussion_r1023953215
########## flink-connectors/flink-connector-files/src/main/java/org/apache/flink/connector/file/table/batch/compact/BatchCompactCoordinator.java: ########## @@ -0,0 +1,155 @@ +/* + * 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.connector.file.table.batch.compact; + +import org.apache.flink.connector.file.table.BinPacking; +import org.apache.flink.connector.file.table.stream.compact.CompactMessages.CompactionUnit; +import org.apache.flink.connector.file.table.stream.compact.CompactMessages.CoordinatorInput; +import org.apache.flink.connector.file.table.stream.compact.CompactMessages.CoordinatorOutput; +import org.apache.flink.connector.file.table.stream.compact.CompactMessages.InputFile; +import org.apache.flink.core.fs.FileSystem; +import org.apache.flink.core.fs.Path; +import org.apache.flink.runtime.state.StateInitializationContext; +import org.apache.flink.streaming.api.operators.AbstractStreamOperator; +import org.apache.flink.streaming.api.operators.BoundedOneInput; +import org.apache.flink.streaming.api.operators.OneInputStreamOperator; +import org.apache.flink.streaming.runtime.streamrecord.StreamRecord; +import org.apache.flink.util.function.SupplierWithException; + +import java.io.IOException; +import java.util.ArrayList; +import java.util.Collections; +import java.util.HashMap; +import java.util.List; +import java.util.Map; +import java.util.function.Function; + +/** + * Coordinator for compaction in batch mode. It will collect the written files in {@link + * BatchFileWriter} and determine whether to compact files or not as well as what files should be + * merged into a single file. + */ +public class BatchCompactCoordinator extends AbstractStreamOperator<CoordinatorOutput> + implements OneInputStreamOperator<CoordinatorInput, CoordinatorOutput>, BoundedOneInput { + + private static final long serialVersionUID = 1L; + + private final SupplierWithException<FileSystem, IOException> fsFactory; + private final long compactAverageSize; + private final long compactTargetSize; + + private transient FileSystem fs; + // the mapping from written partitions to the corresponding files. + private transient Map<String, List<Path>> inputFiles; + + public BatchCompactCoordinator( + SupplierWithException<FileSystem, IOException> fsFactory, + long compactAverageSize, + long compactTargetSize) { + this.fsFactory = fsFactory; + this.compactAverageSize = compactAverageSize; + this.compactTargetSize = compactTargetSize; + } + + @Override + public void initializeState(StateInitializationContext context) throws Exception { + super.initializeState(context); + fs = fsFactory.get(); + inputFiles = new HashMap<>(); + } + + @Override + public void processElement(StreamRecord<CoordinatorInput> element) throws Exception { + CoordinatorInput coordinatorInput = element.getValue(); + if (coordinatorInput instanceof InputFile) { + InputFile file = (InputFile) coordinatorInput; + // collect the written files + inputFiles + .computeIfAbsent(file.getPartition(), k -> new ArrayList<>()) + .add(file.getFile()); + } + } + + @Override + public void endInput() throws Exception { + for (Map.Entry<String, List<Path>> partitionFiles : inputFiles.entrySet()) { + compactPartitionFiles(partitionFiles.getKey(), partitionFiles.getValue()); + } + } + + private void compactPartitionFiles(String partition, List<Path> paths) throws IOException { + int unitId = 0; + final Map<Path, Long> filesSize = getFilesSize(fs, paths); + // calculate the average size of these files + AverageSize averageSize = getAverageSize(filesSize); + if (averageSize.isLessThan(compactAverageSize)) { + // we should compact + // get the written files corresponding to the partition + Function<Path, Long> sizeFunc = filesSize::get; + // determine what files should be merged to a file + List<List<Path>> compactUnits = BinPacking.pack(paths, sizeFunc, compactTargetSize); + for (List<Path> compactUnit : compactUnits) { + // emit the compact units containing the files path + output.collect( + new StreamRecord<>(new CompactionUnit(unitId++, partition, compactUnit))); + } + } else { + // no need to merge these files, emit each single file to downstream for committing + for (Path path : paths) { + output.collect( + new StreamRecord<>( + new CompactionUnit( + unitId++, partition, Collections.singletonList(path)))); + } + } + } + + private Map<Path, Long> getFilesSize(FileSystem fs, List<Path> paths) throws IOException { + Map<Path, Long> filesStatus = new HashMap<>(); + for (Path path : paths) { + long len = fs.getFileStatus(path).getLen(); + filesStatus.put(path, len); + } + return filesStatus; + } + + private AverageSize getAverageSize(Map<Path, Long> filesSize) { + int numFiles = 0; + long totalSz = 0; + for (Map.Entry<Path, Long> fileSize : filesSize.entrySet()) { + numFiles += 1; + totalSz += fileSize.getValue(); + } + return new AverageSize(totalSz, numFiles); + } + + private static class AverageSize { + private final long totalSz; + private final int numFiles; Review Comment: average size can be calculated eagerly -- 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
