org.apache.asterix.external.input.HDFSDataSourceFactory.java Source code

Java tutorial

Introduction

Here is the source code for org.apache.asterix.external.input.HDFSDataSourceFactory.java

Source

/*
 * 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.asterix.external.input;

import java.io.IOException;
import java.util.Arrays;
import java.util.List;
import java.util.Map;

import org.apache.asterix.common.exceptions.AsterixException;
import org.apache.asterix.external.api.AsterixInputStream;
import org.apache.asterix.external.api.IExternalIndexer;
import org.apache.asterix.external.api.IIndexibleExternalDataSource;
import org.apache.asterix.external.api.IRecordReader;
import org.apache.asterix.external.api.IRecordReaderFactory;
import org.apache.asterix.external.indexing.ExternalFile;
import org.apache.asterix.external.indexing.IndexingScheduler;
import org.apache.asterix.external.input.record.reader.IndexingStreamRecordReader;
import org.apache.asterix.external.input.record.reader.hdfs.HDFSRecordReader;
import org.apache.asterix.external.input.record.reader.stream.StreamRecordReader;
import org.apache.asterix.external.input.stream.HDFSInputStream;
import org.apache.asterix.external.provider.ExternalIndexerProvider;
import org.apache.asterix.external.provider.StreamRecordReaderProvider;
import org.apache.asterix.external.provider.StreamRecordReaderProvider.Format;
import org.apache.asterix.external.util.ExternalDataConstants;
import org.apache.asterix.external.util.ExternalDataUtils;
import org.apache.asterix.external.util.HDFSUtils;
import org.apache.hadoop.io.Writable;
import org.apache.hadoop.mapred.InputSplit;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.RecordReader;
import org.apache.hadoop.mapred.Reporter;
import org.apache.hyracks.algebricks.common.constraints.AlgebricksAbsolutePartitionConstraint;
import org.apache.hyracks.api.context.IHyracksTaskContext;
import org.apache.hyracks.api.exceptions.HyracksDataException;
import org.apache.hyracks.hdfs.dataflow.ConfFactory;
import org.apache.hyracks.hdfs.dataflow.InputSplitsFactory;
import org.apache.hyracks.hdfs.scheduler.Scheduler;

public class HDFSDataSourceFactory implements IRecordReaderFactory<Object>, IIndexibleExternalDataSource {

    protected static final long serialVersionUID = 1L;
    protected transient AlgebricksAbsolutePartitionConstraint clusterLocations;
    protected String[] readSchedule;
    protected boolean read[];
    protected InputSplitsFactory inputSplitsFactory;
    protected ConfFactory confFactory;
    protected boolean configured = false;
    protected static Scheduler hdfsScheduler;
    protected static IndexingScheduler indexingScheduler;
    protected static Boolean initialized = false;
    protected static Object initLock = new Object();
    protected List<ExternalFile> files;
    protected Map<String, String> configuration;
    protected Class<?> recordClass;
    protected boolean indexingOp = false;
    private JobConf conf;
    private InputSplit[] inputSplits;
    private String nodeName;
    private Format format;

    @Override
    public void configure(Map<String, String> configuration) throws AsterixException {
        try {
            init();
            this.configuration = configuration;
            JobConf conf = HDFSUtils.configureHDFSJobConf(configuration);
            confFactory = new ConfFactory(conf);
            clusterLocations = getPartitionConstraint();
            int numPartitions = clusterLocations.getLocations().length;
            // if files list was set, we restrict the splits to the list
            InputSplit[] inputSplits;
            if (files == null) {
                inputSplits = conf.getInputFormat().getSplits(conf, numPartitions);
            } else {
                inputSplits = HDFSUtils.getSplits(conf, files);
            }
            if (indexingOp) {
                readSchedule = indexingScheduler.getLocationConstraints(inputSplits);
            } else {
                readSchedule = hdfsScheduler.getLocationConstraints(inputSplits);
            }
            inputSplitsFactory = new InputSplitsFactory(inputSplits);
            read = new boolean[readSchedule.length];
            Arrays.fill(read, false);
            String formatString = configuration.get(ExternalDataConstants.KEY_FORMAT);
            if (formatString == null || formatString.equals(ExternalDataConstants.FORMAT_HDFS_WRITABLE)) {
                RecordReader<?, ?> reader = conf.getInputFormat().getRecordReader(inputSplits[0], conf,
                        Reporter.NULL);
                this.recordClass = reader.createValue().getClass();
                reader.close();
            } else {
                format = StreamRecordReaderProvider.getReaderFormat(configuration);
                this.recordClass = char[].class;
            }
        } catch (IOException e) {
            throw new AsterixException(e);
        }
    }

    // Used to tell the factory to restrict the splits to the intersection between this list a
    // actual files on hde
    @Override
    public void setSnapshot(List<ExternalFile> files, boolean indexingOp) {
        this.files = files;
        this.indexingOp = indexingOp;
    }

    /*
     * The method below was modified to take care of the following
     * 1. when target files are not null, it generates a file aware input stream that validate
     * against the files
     * 2. if the data is binary, it returns a generic reader */
    public AsterixInputStream createInputStream(IHyracksTaskContext ctx, int partition, IExternalIndexer indexer)
            throws HyracksDataException {
        try {
            if (!configured) {
                conf = confFactory.getConf();
                inputSplits = inputSplitsFactory.getSplits();
                nodeName = ctx.getJobletContext().getApplicationContext().getNodeId();
                configured = true;
            }
            return new HDFSInputStream(read, inputSplits, readSchedule, nodeName, conf, configuration, files,
                    indexer);
        } catch (Exception e) {
            throw new HyracksDataException(e);
        }
    }

    /**
     * Get the cluster locations for this input stream factory. This method specifies on which asterix nodes the
     * external
     * adapter will run and how many threads per node.
     *
     * @return
     */
    @Override
    public AlgebricksAbsolutePartitionConstraint getPartitionConstraint() {
        clusterLocations = HDFSUtils.getPartitionConstraints(clusterLocations);
        return clusterLocations;
    }

    /**
     * This method initialize the scheduler which assigns responsibility of reading different logical input splits from
     * HDFS
     */
    private static void init() {
        if (!initialized) {
            synchronized (initLock) {
                if (!initialized) {
                    hdfsScheduler = HDFSUtils.initializeHDFSScheduler();
                    indexingScheduler = HDFSUtils.initializeIndexingHDFSScheduler();
                    initialized = true;
                }
            }
        }
    }

    public JobConf getJobConf() throws HyracksDataException {
        return confFactory.getConf();
    }

    @Override
    public DataSourceType getDataSourceType() {
        return ExternalDataUtils.getDataSourceType(configuration);
    }

    /**
     * HDFS Datasource is a special case in two ways:
     * 1. It supports indexing.
     * 2. It returns input as a set of writable object that we sometimes internally transform into a byte stream
     * Hence, it can produce:
     * 1. StreamRecordReader: When we transform the input into a byte stream.
     * 2. Indexing Stream Record Reader: When we transform the input into a byte stream and perform indexing.
     * 3. HDFS Record Reader: When we simply pass the Writable object as it is to the parser.
     */
    @Override
    public IRecordReader<? extends Object> createRecordReader(IHyracksTaskContext ctx, int partition)
            throws HyracksDataException {
        try {
            IExternalIndexer indexer = files == null ? null : ExternalIndexerProvider.getIndexer(configuration);
            if (format != null) {
                StreamRecordReader streamReader = StreamRecordReaderProvider.createRecordReader(format,
                        createInputStream(ctx, partition, indexer), configuration);
                if (indexer != null) {
                    return new IndexingStreamRecordReader(streamReader, indexer);
                } else {
                    return streamReader;
                }
            }
            JobConf conf = confFactory.getConf();
            InputSplit[] inputSplits = inputSplitsFactory.getSplits();
            String nodeName = ctx.getJobletContext().getApplicationContext().getNodeId();
            return new HDFSRecordReader<Object, Writable>(read, inputSplits, readSchedule, nodeName, conf, files,
                    indexer);
        } catch (Exception e) {
            throw new HyracksDataException(e);
        }
    }

    @Override
    public Class<?> getRecordClass() {
        return recordClass;
    }

    @Override
    public boolean isIndexible() {
        return true;
    }

    @Override
    public boolean isIndexingOp() {
        return ((files != null) && indexingOp);
    }
}