org.apache.avro.mapred.AvroOutputFormat Maven / Gradle / Ivy
/**
* 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.avro.mapred;
import java.io.IOException;
import java.io.UnsupportedEncodingException;
import java.util.Map;
import java.net.URLDecoder;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.Reporter;
import org.apache.hadoop.mapred.FileOutputFormat;
import org.apache.hadoop.mapred.RecordWriter;
import org.apache.hadoop.util.Progressable;
import org.apache.avro.Schema;
import org.apache.avro.file.DataFileWriter;
import org.apache.avro.file.CodecFactory;
import org.apache.avro.generic.GenericData;
import org.apache.avro.hadoop.file.HadoopCodecFactory;
import static org.apache.avro.file.DataFileConstants.DEFAULT_SYNC_INTERVAL;
import static org.apache.avro.file.DataFileConstants.DEFLATE_CODEC;
import static org.apache.avro.file.DataFileConstants.XZ_CODEC;
import static org.apache.avro.file.CodecFactory.DEFAULT_DEFLATE_LEVEL;
import static org.apache.avro.file.CodecFactory.DEFAULT_XZ_LEVEL;
/**
* An {@link org.apache.hadoop.mapred.OutputFormat} for Avro data files.
*
* You can specify various options using Job Configuration properties.
* Look at the fields in {@link AvroJob} as well as this class to get
* an overview of the supported options.
*/
public class AvroOutputFormat
extends FileOutputFormat, NullWritable> {
/** The file name extension for avro data files. */
public final static String EXT = ".avro";
/** The configuration key for Avro deflate level. */
public static final String DEFLATE_LEVEL_KEY = "avro.mapred.deflate.level";
/** The configuration key for Avro XZ level. */
public static final String XZ_LEVEL_KEY = "avro.mapred.xz.level";
/** The configuration key for Avro sync interval. */
public static final String SYNC_INTERVAL_KEY = "avro.mapred.sync.interval";
/** Enable output compression using the deflate codec and specify its level.*/
public static void setDeflateLevel(JobConf job, int level) {
FileOutputFormat.setCompressOutput(job, true);
job.setInt(DEFLATE_LEVEL_KEY, level);
}
/** Set the sync interval to be used by the underlying {@link DataFileWriter}.*/
public static void setSyncInterval(JobConf job, int syncIntervalInBytes) {
job.setInt(SYNC_INTERVAL_KEY, syncIntervalInBytes);
}
static void configureDataFileWriter(DataFileWriter writer,
JobConf job) throws UnsupportedEncodingException {
CodecFactory factory = getCodecFactory(job);
if (factory != null) {
writer.setCodec(factory);
}
writer.setSyncInterval(job.getInt(SYNC_INTERVAL_KEY, DEFAULT_SYNC_INTERVAL));
// copy metadata from job
for (Map.Entry e : job) {
if (e.getKey().startsWith(AvroJob.TEXT_PREFIX))
writer.setMeta(e.getKey().substring(AvroJob.TEXT_PREFIX.length()),
e.getValue());
if (e.getKey().startsWith(AvroJob.BINARY_PREFIX))
writer.setMeta(e.getKey().substring(AvroJob.BINARY_PREFIX.length()),
URLDecoder.decode(e.getValue(), "ISO-8859-1")
.getBytes("ISO-8859-1"));
}
}
/** This will select the correct compression codec from the JobConf.
* The order of selection is as follows:
*
* - If mapred.output.compress is true then look for codec otherwise no compression
* - Use avro.output.codec if populated
* - Next use mapred.output.compression.codec if populated
* - If not default to Deflate Codec
*
*/
static CodecFactory getCodecFactory(JobConf job) {
CodecFactory factory = null;
if (FileOutputFormat.getCompressOutput(job)) {
int deflateLevel = job.getInt(DEFLATE_LEVEL_KEY, DEFAULT_DEFLATE_LEVEL);
int xzLevel = job.getInt(XZ_LEVEL_KEY, DEFAULT_XZ_LEVEL);
String codecName = job.get(AvroJob.OUTPUT_CODEC);
if (codecName == null) {
String codecClassName = job.get("mapred.output.compression.codec", null);
String avroCodecName = HadoopCodecFactory.getAvroCodecName(codecClassName);
if ( codecClassName != null && avroCodecName != null){
factory = HadoopCodecFactory.fromHadoopString(codecClassName);
job.set(AvroJob.OUTPUT_CODEC , avroCodecName);
return factory;
} else {
return CodecFactory.deflateCodec(deflateLevel);
}
} else {
if ( codecName.equals(DEFLATE_CODEC)) {
factory = CodecFactory.deflateCodec(deflateLevel);
} else if ( codecName.equals(XZ_CODEC)) {
factory = CodecFactory.xzCodec(xzLevel);
} else {
factory = CodecFactory.fromString(codecName);
}
}
}
return factory;
}
@Override
public RecordWriter, NullWritable>
getRecordWriter(FileSystem ignore, JobConf job,
String name, Progressable prog)
throws IOException {
boolean isMapOnly = job.getNumReduceTasks() == 0;
Schema schema = isMapOnly
? AvroJob.getMapOutputSchema(job)
: AvroJob.getOutputSchema(job);
GenericData dataModel = AvroJob.createDataModel(job);
final DataFileWriter writer =
new DataFileWriter(dataModel.createDatumWriter(null));
configureDataFileWriter(writer, job);
Path path = FileOutputFormat.getTaskOutputPath(job, name+EXT);
writer.create(schema, path.getFileSystem(job).create(path));
return new RecordWriter, NullWritable>() {
public void write(AvroWrapper wrapper, NullWritable ignore)
throws IOException {
writer.append(wrapper.datum());
}
public void close(Reporter reporter) throws IOException {
writer.close();
}
};
}
}