com.blackberry.logdriver.mapreduce.avro.AvroBlockInputFormat.java Source code

Introduction

Here is the source code for com.blackberry.logdriver.mapreduce.avro.AvroBlockInputFormat.java

Source

/** Copyright (c) 2014 BlackBerry Limited
 *
 *  Licensed 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. 
 */

/**
 * Reads in whole blocks of Avro files without decompressing or deserializing
 *  them.
 * 
 * There are two uses for this.
 * 
 * <h3>Merging Multiple Avro Files</h3>
 * To merge multiple Avro files that have the same schema and codec, read in
 * multiple files with the AvroBlockInputFormat, map them with
 * AvroBlockWriterMapper (to remove all but one header, and clean up the sync
 * markers), and write the result out with BinaryOutputFormat.
 * 
 * <h3>Low Level Optimizations</h3>
 * Maybe you only want to process certain blocks, and maybe you can tell
 * which blocks those are without deserializing the data.  In that case, you
 * can filter the raw blocks and only process certain ones.  Since
 * deserialization can be a relatively expensive operation, this might be 
 * useful in certain situations.
 * 
 */
package com.blackberry.logdriver.mapreduce.avro;

import java.io.IOException;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.BytesWritable;
import org.apache.hadoop.mapreduce.InputSplit;
import org.apache.hadoop.mapreduce.JobContext;
import org.apache.hadoop.mapreduce.RecordReader;
import org.apache.hadoop.mapreduce.TaskAttemptContext;
import org.apache.hadoop.mapreduce.lib.input.CombineFileInputFormat;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import com.blackberry.logdriver.avro.AvroFileHeader;

public class AvroBlockInputFormat extends CombineFileInputFormat<AvroFileHeader, BytesWritable> {
    private static final Logger LOG = LoggerFactory.getLogger(AvroBlockInputFormat.class);
    private static final long MAX_SPLIT_LOCATIONS = 100000;

    /**
     * Avro files are always splitable.
     * 
     * @param context
     *          The JobContext.
     * @param file
     *          The file which is being checked. Since this is an Avro input
     *          format, we assume it is an Avro file.
     * @return true
     */
    @Override
    protected boolean isSplitable(JobContext context, Path file) {
        return true;
    }

    /**
     * Creates a new AvroBlockRecordReader.
     * 
     * Increases there default value mapreduce.job.max.split.locations to 100000,
     * if it's not already set.
     * 
     * Also sets mapred.max.split.size to the default block size for the root
     * directory ("/"), if it's not already set.
     * 
     * @param split
     *          The InputSplit.
     * @param context
     *          The TaskAttemptContext.
     * @return A new AvroBlockRecordReader.
     * @throws IOException
     *           If there is an I/O error.
     */
    @SuppressWarnings("deprecation")
    @Override
    public RecordReader<AvroFileHeader, BytesWritable> createRecordReader(InputSplit split,
            TaskAttemptContext context) throws IOException {
        Configuration conf = context.getConfiguration();

        // Ensure we have sensible defaults for how we build blocks.
        if (conf.get("mapreduce.job.max.split.locations") == null) {
            conf.setLong("mapreduce.job.max.split.locations", MAX_SPLIT_LOCATIONS);
        }
        if (conf.get("mapred.max.split.size") == null) {
            // Try to set the split size to the default block size. In case of
            // failure, we'll use this 128MB default.
            long blockSize = 128 * 1024 * 1024; // 128MB
            try {
                blockSize = FileSystem.get(conf).getDefaultBlockSize();
            } catch (IOException e) {
                LOG.error("Error getting filesystem to get get default block size (this does not bode well).");
            }
            conf.setLong("mapred.max.split.size", blockSize);
        }

        return new AvroBlockRecordReader();
    }

}