Hadoop源码阅读之WritableComparator类

1. 类简介

  • All Implemented Interfaces: Comparator, Configurable, RawComparator

  • 包名：org.apache.hadoop.io.WritableComparator

2. 类释义

A Comparator for WritableComparables.

  • WritableComparables 的比较器

This base implemenation uses the natural ordering. To define alternate orderings, override compare(WritableComparable).

  • 使用natural ordering的基本实现。 如果要实现自定义排序，那么请覆写compare()【来自WritbaleComparable类的方法】

One may optimize compare-intensive operations by overriding compare(byte[],int,int,byte[],int,int). Static utility methods are provided to assist in optimized implementations of this method.

  • 可以通过重写compare(byte[],int,int,byte[],int,int)来优化比较密集型操作。提供静态实用程序方法来帮助优化此方法的实现。

3. 类方法

  • compare方法
    compare方法有好几种不同的实现，以下逐一讲解：
public int compare(WritableComparable a,
          WritableComparable b)

Compare two WritableComparables.

The default implementation uses the natural ordering, calling Comparable.compareTo(Object).

public int compare(byte[] b1,
          int s1,
          int l1,
          byte[] b2,
          int s2,
          int l2)

Optimization hook. Override this to make SequenceFile.Sorter's scream.

The default implementation reads the data into two WritableComparables (using Writable.readFields(DataInput), then calls compare(WritableComparable,WritableComparable).

Specified by:
    compare in interface RawComparator
Parameters:
    b1 - The first byte array.
    s1 - The position index in b1. The object under comparison's starting index.
    l1 - The length of the object in b1.
    b2 - The second byte array.
    s2 - The position index in b2. The object under comparison's starting index.
    l2 - The length of the object under comparison in b2.
Returns:
    An integer result of the comparison.

public int compare(Object a,
          Object b)

Specified by:
    compare in interface Comparator