1 需求
无论 hdfs 还是 mapreduce,对于小文件都有损效率,实践中,又难免面临处理大量小文件的场景,此时,就需要有相应解决方案。 将多个小文件合并成一个文件 SequenceFile,SequenceFile 里面存储着多个文件,存储的形式为文件路径+名称为 key, 文件内容为 value。
2 分析
小文件的优化无非以下几种方式:
1)在数据采集的时候,就将小文件或小批数据合成大文件再上传 HDFS
2) 在业务处理之前, 在 HDFS 上使用 mapreduce 程序对小文件进行合并
3)在 mapreduce 处理时,可采用 CombineTextInputFormat 提高效率
3 具体实现
本节采用自定义 InputFormat 的方式, 处理输入小文件的问题。
1)自定义一个类继承 FileInputFormat
2)改写 RecordReader,实现一次读取一个完整文件封装为 KV
3)在输出时使用 SequenceFileOutPutFormat 输出合并文件
4)程序实现:
1)自定义 InputFromat
package com.da.inputformat;
import java.io.IOException;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.BytesWritable;
import org.apache.hadoop.io.NullWritable;
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.FileInputFormat;
public class WholeFileInputformat extends FileInputFormat<NullWritable, BytesWritable> {
@Override
protected boolean isSplitable(JobContext context, Path filename) {
// 不能切割
return false;
}
@Override
public RecordReader<NullWritable, BytesWritable> createRecordReader(InputSplit split, TaskAttemptContext context)
throws IOException, InterruptedException {
// 创建对象
WholeRecordReader recordReader = new WholeRecordReader();
// 初始化
recordReader.initialize(split, context);
return recordReader;
}
}
2) 自定义 RecordReader
package com.da.inputformat;
import java.io.IOException;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FSDataInputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.BytesWritable;
import org.apache.hadoop.io.IOUtils;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.mapreduce.InputSplit;
import org.apache.hadoop.mapreduce.RecordReader;
import org.apache.hadoop.mapreduce.TaskAttemptContext;
import org.apache.hadoop.mapreduce.lib.input.FileSplit;
public class WholeRecordReader extends RecordReader<NullWritable, BytesWritable> {
private BytesWritable value = new BytesWritable();
private boolean isProgress = false;
private FileSplit split;
private Configuration configuration;
@Override
public void initialize(InputSplit split, TaskAttemptContext context) throws IOException, InterruptedException {
// 初始化
this.split = (FileSplit) split;
this.configuration = context.getConfiguration();
}
@Override
public boolean nextKeyValue() throws IOException, InterruptedException {
// 读取一个个的文件
if (!isProgress) {
// 1缓冲区
byte buf[] = new byte[(int) split.getLength()];
FileSystem fs = null;
FSDataInputStream fis = null;
try {
// 2获取文件系统
Path path = split.getPath();
fs = path.getFileSystem(configuration);
// 3打开文件输入流
fis = fs.open(path);
// 4流的拷贝
IOUtils.readFully(fis, buf, 0, buf.length);
// 5 输出文件内容
value.set(buf, 0, buf.length);
} catch (Exception e) {
} finally {
IOUtils.closeStream(fis);
IOUtils.closeStream(fs);
}
isProgress = true;
return true;
}
return false;
}
@Override
public NullWritable getCurrentKey() throws IOException, InterruptedException {
return NullWritable.get();
}
@Override
public BytesWritable getCurrentValue() throws IOException, InterruptedException {
return value;
}
@Override
public float getProgress() throws IOException, InterruptedException {
// 当前进度
return isProgress ? 1 : 0;
}
@Override
public void close() throws IOException {
}
}
3) SequenceFileMapper 处理流程
package com.da.inputformat;
import java.io.IOException;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.BytesWritable;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.lib.input.FileSplit;
public class SequenceFileMapper extends Mapper<NullWritable, BytesWritable, Text, BytesWritable> {
private Text k = new Text();
@Override
protected void setup(Mapper<NullWritable, BytesWritable, Text, BytesWritable>.Context context)
throws IOException, InterruptedException {
// 获取文件的路径与名称
FileSplit split = (FileSplit) context.getInputSplit();
Path path = split.getPath();
k.set(path.toString());
}
@Override
protected void map(NullWritable key, BytesWritable value, Context context)
throws IOException, InterruptedException {
context.write(k, value);
}
}
4) SequenceFileReducer 处理流程
package com.da.inputformat;
import java.io.IOException;
import org.apache.hadoop.io.BytesWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Reducer;
public class SequenceFileReducer extends Reducer<Text, BytesWritable, Text, BytesWritable> {
@Override
protected void reduce(Text key, Iterable<BytesWritable> values, Context context)
throws IOException, InterruptedException {
for (BytesWritable bytesWritable : values) {
context.write(key, bytesWritable);
}
}
}
5) SequenceFileDriver 处理流程
package com.da.inputformat;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.BytesWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.mapreduce.lib.output.SequenceFileOutputFormat;
public class SequenceFileDriver {
public static void main(String[] args) throws Exception {
args = new String[] { "E:/mrinput/", "E:/mrout" };
Configuration conf = new Configuration();
Job job = Job.getInstance(conf);
job.setJarByClass(SequenceFileDriver.class);
job.setMapperClass(SequenceFileMapper.class);
job.setReducerClass(SequenceFileReducer.class);
// 设置输入的 inputFormat
job.setInputFormatClass(WholeFileInputformat.class);
// 设置输出的 outputFormat
job.setOutputFormatClass(SequenceFileOutputFormat.class);
job.setMapOutputKeyClass(Text.class);
job.setMapOutputValueClass(BytesWritable.class);
job.setOutputKeyClass(Text.class);
job.setOutputValueClass(BytesWritable.class);
FileInputFormat.setInputPaths(job, new Path(args[0]));
FileOutputFormat.setOutputPath(job, new Path(args[1]));
boolean result = job.waitForCompletion(true);
System.exit(result ? 0 : 1);
}
}