这篇文章主要讲解了“Hadoop 序列化怎么实现”,文中的讲解内容简单清晰,易于学习与理解,下面请大家跟着丸趣 TV 小编的思路慢慢深入,一起来研究和学习“Hadoop 序列化怎么实现”吧!
Hadoop I/OData Integrity
Hdfs: % hadoop fs -cat hdfs://namenode/data/a.txt
LocalFS: % hadoop fs -cat file:///tmp/a.txt
generate crc check sum file
%hadoop fs -copyToLocal -crc /data/a.txt file:///data/a.txt
check sum file: .a.txt.crc is a hidden file.
Ref: CRC-32, 循环冗余校验算法,error-detecting.
io.bytes.per.checksum is deprecated, it s dfs.bytes-per-checksum, default is 512, Must not be larger than dfs.stream-buffer-size,which is the size of buffer to stream files. The size of this buffer should probably be a multiple of hardware page size (4096 on Intel x86), and it determines how much data is buffered during read and write operations.
Data Compression
常用算法
读书时,hadoop 支持四种压缩算法, 如果调解空间和效率的话,-1 ~ -9, 代表从最优速度到最优空间. 压缩算法支持在 org.apache.hadoop.io.compress.*.
deflate (.deflate), 就是常用的 gzip, package ..DefaultCodec
Gzip (.gz), 在 deflate 格式加了文件头和尾. 压缩速度 (适中), 解压速度 (适中), 压缩效率 (适中),package ..GzipCodec, both of java and native
bzip2 (.bz2), 压缩速度 (最差), 解压速度 (最差), 压缩效率 (最好),特点是支持可切分 (splitable),对 map-red 非常友好。,package ..BZip2Codec,java only
LZO (.lzo), 压缩速度 (最快), 解压速度 (最快), 压缩效率 (最差),,package com.hadoop.compressiojn.lzo.lzopCodec, native only
如果禁用原生库, 使用 hadoop.native.lib.
如果使用原生库, 可能对象创建的成本较高, 所以可以使用 CodecPool,重复使用这些对象。
对于一个非常大的数据文件,存储如下方案:
使用支持切分的 bzip2
手动切分,并使压缩后的 part 接近于 block size.
使用 Sequence File, 它支持压缩和切分
使用 Avro 数据文件,它也支持压缩和切分,而且增加了很多编程语言的可读写性。
如果 Map-Red 的 output 自动压缩:
conf.setBoolean (mared.output.compress ,true);
conf.setClass(mapred.output.compression.codec ,GzipCodec.class,CompressionCodec.class);
如果 Map-Red 的中间结果的自动压缩:
//or conf.setCompressMapOutput(true);
conf.setBoolean (mared.compress.map.output ,true);
//or conf.setMapOutputComressorClass(GzipCodec.class)
conf.setClass(mapred.map.output.compression.codec ,GzipCodec.class,CompressionCodec.class);
序列化 (Serialization/Deserialization)Writable and WritableComparable
// core class for hadoop
public interface Writable{ void write(DataOutput out) throw IOException;
void readFields(DataInput in) throw IOException;
public interface Comparable T { int compareTo(T o);
//core class for map-reduce shuffle
public interface WritableComparable T extends Writable, Comparable T {
// Sample
public class MyWritableComparable implements WritableComparable {
// Some data
private int counter;
private long timestamp;
public void write(DataOutput out) throws IOException { out.writeInt(counter);
out.writeLong(timestamp);
}
public void readFields(DataInput in) throws IOException { counter = in.readInt();
timestamp = in.readLong();
}
public int compareTo(MyWritableComparable o) {
int thisValue = this.value;
int thatValue = o.value;
return (thisValue thatValue ? -1 : (thisValue==thatValue ? 0 : 1));
}
public int hashCode() {
final int prime = 31;
int result = 1;
result = prime * result + counter;
result = prime * result + (int) (timestamp ^ (timestamp 32));
return result
}
//optimize for stream comparasion
public interface RawComparator T extends Comparator T {
// s1 start position, l1, length of bytes
public int compare(byte[] b1, int s1,int l1,byte[] b2,int s2,int l2);
public class WritableComparator implements RawComparator{}
Comparator RawComparator WritableComparator
WritableComparator 提供了原始 compator 的 compare 反序列化对象的实现,性能较差。不过它作为 RawComparator 实例的工厂:
RawComparator IntWritable comparator = WritableComparator.get(IntWritable.class);
// 注册一个经过优化的比较算子。Register an optimized comparator for a WritableComparable implementation.
static void define(Class c, WritableComparator comparator);
// 获得一个 WritableComparable 的比较算子. Get a comparator for a WritableComparable implementation.
static WritableComparator get(Class ? extends WritableComparable
public MyWritableComparator extends WritableComparator{
static{ define(MyWritableComparable.class, new MyWritableComparator());
}
public MyWritableComparator { super(MyWritableComparable.class);
}
@Override
public int compare(byte[] b1, int s1,int l1,byte[] b2,int s2,int l2){ }
}
注: 要使 static initializer 被调用,除非有该类的实例被创建,或某静态方法或成员被访问。或者直接强制,代码如:
Class.forName(package.yourclass 它会强制初始化静态 initializer.
Java Primitive Data Type wrapped by WritableExtends from WritableComparable
BooleanWritable, 1
ByteWritable, 1,
BytesWritable,
IntWritable,4
VIntWritable,1~5
FloatWritable,4,
LongWritable,8,
VLongWritable,1~9
DoubleWritable,8
NullWritable,Immutable singletone.
Text,4~
MD5Hash,
ObjectWritable,
GenericWritable
Extends from Writable only
ArrayWritable
TwoDArrayWritable
AbstractMapWritable
MapWritable
SortedMapWritable
[Text]
值得一提的是 Text 的序列化方式是 Zero-compressed encoding, 这个看过一些资料,其实是一种编码方式,意图是省略掉高位 0 所占用的空间,对于小数,它能节省空间,对于大数会额外占用空间。相比压缩,它能比较快速。其实类似于 VIntWritable, VLongWritable 的编码方式。
– 如何选择变长和定长数值呢?
1. 定长适合分布非常均匀的数值(如 hash),变长适合分布非常不均匀的数值。
2. 变长可以节省空间,而且可以在 VIntWritable 和 VLongWritable 之间转换。
– Text 和 String 的区别
1。String 是 char 序列,Text 是 UTF- 8 的 byte 序列.
UTF- 8 类不能对字符串大于 32767 的进行 utf- 8 编码。
(Indexing) 索引:对于 ASCII 来说,Text 和 String 是一样的,对于 Unicode 就不同了。String 类的长度是其所含 char 编码单元的长度,然而 Text 是 UTF- 8 的字节码的长度。CodePointAt 表示一个真正的 Unicode 字符,它可以是 2char,4bytes 的 unicode。
Iteration(迭代): 将 Text 转换 ByteBuffer, 然后反复调用 bytesToCodePoint() 静态方法,可以取到整型的 Unicode.
Mutable(易变性): 可以 set,类似 writable 和 StringBuffer,getLength() 返回有效字串长度,getbytes().length, 返回空间大小。
[BytesWritable]
这是二进制数组的封装,类似于 windows 下的 BSTR,都是前面一个整型表示字节长度,后面是字节的二进制流。
它也是 mutable,getLength() != getBytes().length
[NullWritable]
NullWritable 是 Writable 的一个特殊类型。它的序列化长度为 0,其实只是一个占位符,既不读入,也不写出。只是存在于程序体中。
Immutable, 是一个 singleton。
[ObjectWritable]
ObjectWritable 是 Java 的 Array, String, 以及 Primitive 类型的通用封装 (注:不包含 Integer)。它的序列化则使用 java 的类型序列化,写入类型信息等,比较占用空间。
通过两个特殊的构造:
public ObjectWritable(Object instance);
public ObjectWritable(Class declaredClass,Object instance);
举例子:
ObjectWritable objectw = new ObjectWritable(int.class,5);
[GenericWritable]
首先这是一个抽象类,需要被具象化才能使用。
观察下面这个实列,它以一种 Union 方式,显示的代理一个 Writable 实例,解决了 Reduce 函数的参数声明问题。
public class MyGenericWritable extends GenericWritable { private static Class ? extends Writable [] CLASSES = null;
static { CLASSES = (Class ? extends Writable []) new Class[] {
IntWritable.class,
Text.class
//add as many different Writable class as you want
};
}
@Override
protected Class ? extends Writable [] getTypes() {
return CLASSES;
}
@Override
public String toString() { return MyGenericWritable [getTypes()= + Arrays.toString(getTypes()) + ]
}
// override hashcode();
public class Reduce extends Reducer Text, MyGenericWritable, Text, Text { public void reduce(Text key, Iterable MyGenericWritable values, Context context) throws IOException, InterruptedException {}
[ArrayWritable /TwoDArrayWritable]
ArrayWritable aw = new ArrayWriable(Text.class);
[MapWritable / SortedMapWritable]
实现了 java.util.Map Writable,Writable 和 SortedMap…
它的 serialize, 使用先写 map classname,id , 然后后边每个类的类型,以 id 来替代,节省空间。这些都在父类 AbstractMapWritable 中实现。
集合小结:
1. 如果是单类型的列表, 使用 ArrayWritable 就足够了
2。如果是把不同类型的 Writable 存储在一个列表中:
— 可以使用 GenerickWritable, 把元素封装在一个 ArrayWritable,这个貌似只能同一类型。
public class MyGenericWritable extends GenericWritable { private static Class ? extends Writable [] CLASSES = null;
static { CLASSES = (Class ? extends Writable []) new Class[] {
ArrayWritable.class,
//add as many different Writable class as you want
};
}
@Override
protected Class ? extends Writable [] getTypes() {
return CLASSES;
}
— 可以使用写一个仿照 MapWritable 的 ListWritable
// 注意实现 hashcode,equals,toString, comparTo (if possible)
//hashcode 尤其重要,HashPartitioner 通常用 hashcode 来选择 reduce 分区,所以为你的类写一个比较好的 hashcode 非常必要。
public class ListWritable extends ArrayList Writable implements Writable {
}
/**
* @author cloudera
*
*/
public class ListWritable extends ArrayList Writable implements Writable {
private List Writable list = new ArrayList Writable
public void set(Writable writable){list.add(writable);
@Override
public void readFields(DataInput in) throws IOException {int nsize = in.readInt();
Configuration conf = new Configuration();
Text className = new Text();
while(nsize-- 0){
Class theClass = null;
try {className.readFields(in);
theClass = Class.forName(className.toString());
} catch (ClassNotFoundException e) {
// TODO Auto-generated catch block
e.printStackTrace();
Writable w = (Writable)ReflectionUtils.newInstance(theClass,conf);
w.readFields(in);
add(w);
@Override
public void write(DataOutput out) throws IOException {
Writable w = null;
out.writeInt(size());
for(int i = 0;i size();i++){w = get(i);
new Text(w.getClass().getName()).write(out);
w.write(out);
}
感谢各位的阅读,以上就是“Hadoop 序列化怎么实现”的内容了,经过本文的学习后,相信大家对 Hadoop 序列化怎么实现这一问题有了更深刻的体会,具体使用情况还需要大家实践验证。这里是丸趣 TV,丸趣 TV 小编将为大家推送更多相关知识点的文章,欢迎关注!