/*
* 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.jena.hadoop.rdf.io.output.writers;
import java.io.IOException;
import java.io.Writer;
import org.apache.hadoop.mapreduce.RecordWriter;
import org.apache.hadoop.mapreduce.TaskAttemptContext;
import org.apache.jena.hadoop.rdf.types.AbstractNodeTupleWritable;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
/**
* An abstract implementation of a record writer that writes records in batches.
* <p>
* It is important to note that the writer will write output periodically once
* sufficient tuples have been gathered. If there is an incomplete batch when
* the {@link #close(TaskAttemptContext)} method is called then the final batch
* will be written then. Writing in batches increases the chances that the
* writer will be able to effectively use the syntax compressions of the RDF
* serialization being used.
* </p>
* <p>
* The implementation only writes the value portion of the key value pair since
* it is the value portion that is used to convey the node tuples
* </p>
*
*
*
* @param <TKey>
* @param <TValue>
* @param <T>
*/
public abstract class AbstractBatchedNodeTupleWriter<TKey, TValue, T extends AbstractNodeTupleWritable<TValue>> extends
RecordWriter<TKey, T> {
private static final Logger LOG = LoggerFactory.getLogger(AbstractBatchedNodeTupleWriter.class);
private Writer writer;
private long batchSize;
protected AbstractBatchedNodeTupleWriter(Writer writer, long batchSize) {
if (writer == null)
throw new NullPointerException("writer cannot be null");
if (batchSize <= 0)
throw new IllegalArgumentException("batchSize must be >= 1");
this.writer = writer;
this.batchSize = batchSize;
}
@Override
public final void write(TKey key, T value) throws IOException {
LOG.debug("write({}={})", key, value);
if (this.add(value) >= this.batchSize) {
long size = this.writeOutput(writer);
if (size > 0)
throw new IOException("Derived implementation failed to empty the current batch after writing");
}
}
/**
* Adds the tuple to the batch of tuples that will be written when the batch
* threshold is reached or when the {@link #close(TaskAttemptContext)}
* method is called.
*
* @param value
* Tuple
* @return The current size of the batch waiting to be written
*/
protected abstract long add(T value);
@Override
public void close(TaskAttemptContext context) throws IOException {
if (this.writer != null) {
long size = this.writeOutput(writer);
if (size > 0)
throw new IOException("Derived implementation failed to empty the current batch after writing");
this.writer.close();
this.writer = null;
}
}
/**
* Writes the current batch of tuples to the writer, the writer should not
* be closed and the batch should be emptied by the implementation.
* <p>
* If the current batch is empty then this should be a no-op
* </p>
*
* @param writer
* Writer
* @return Current batch size which should always be zero
*/
protected abstract long writeOutput(Writer writer);
}