hadoop自定义key，value

2019独角兽企业重金招聘Python工程师标准>>>

Hadoop的自定制数据类型
一般有两个办法，一种较为简单的是针对值，另外一种更为完整的是对于键和值都适应的方法：
1、实现Writable接口：

/* DataInput and DataOutput 类是java.io的类 */
public interface Writable {
    void readFields(DataInput in);
    void write(DataOutput out);
}

下面是一个小例子：
public class Point3D implement Writable {
 public float x, y, z;
 public Point3D(float fx, float fy, float fz) {
  this.x = fx;
  this.y = fy;
  this.z = fz;
 }
 public Point3D() {
  this(0.0f, 0.0f, 0.0f);
 }
 public void readFields(DataInput in) throws IOException {
  x = in.readFloat();
  y = in.readFloat();
  z = in.readFloat();
 }
 public void write(DataOutput out) throws IOException {
  out.writeFloat(x);
  out.writeFloat(y);
  out.writeFloat(z);
 }
 public String toString() {
  return Float.toString(x) + ", "
   + Float.toString(y) + ", "
   + Float.toString(z);
 }
}

2、对于键来说，需要指定排序规则，对此，Java版Hadoop的办法是实现WritableComparable这个泛型接口，WritableComparable，顾名思义了，一半是Writable，一半是Comparable

public interface WritableComparable<T> {
 public void readFields(DataInput in);
 public void write(DataOutput out);
 public int compareTo(T other);
}

这里的compareTo方法是默认的key排序

先给出下面的简单例子，再做说明和扩展。
public class Point3D inplements WritableComparable {
 public float x, y, z;
 public Point3D(float fx, float fy, float fz) {
  this.x = fx;
  this.y = fy;
  this.z = fz;
 }
 public Point3D() {
  this(0.0f, 0.0f, 0.0f);
 }
 public void readFields(DataInput in) throws IOException {
  x = in.readFloat();
  y = in.readFloat();
  z = in.readFloat();
 }
 public void write(DataOutput out) throws IOException {
  out.writeFloat(x);
  out.writeFloat(y);
  out.writeFloat(z);
 }
 public String toString() {
  return Float.toString(x) + ", "
   + Float.toString(y) + ", "
   + Float.toString(z);
 }
 public float distanceFromOrigin() {
  return (float) Math.sqrt( x*x + y*y +z*z);
 }
 public int compareTo(Point3D other) {
  return Float.compareTo(
   distanceFromOrigin(),
   other.distanceFromOrigin());
 }
 public boolean equals(Object o) {
  if( !(o instanceof Point3D)) {
   return false;
  }
  Point3D other = (Point3D) o;
  return this.x == o.x
   && this.y == o.y
   && this.z == o.z;
 }
 /* 实现 hashCode() 方法很重要
  * Hadoop的Partitioners会用到这个方法，后面再说
  */
 public int hashCode() {
  return Float.floatToIntBits(x)
   ^ Float.floatToIntBits(y)
   ^ Float.floatToIntBits(z);
 }
}
如果要将对象写入数据库则还要继承DBWritable接口

public interface WritableComparable<T> {
 public void write(PreparedStatement statement) throwsSQLException;
public void readFields(ResultSet resultSet) throws SQLException;}

下面写个例子

public class LocationBean implements Writable, DBWritable { 
    private String mobilenetworkcode; 
 
    private String mobilecountrycode; 
 
    private Integer cellid; 
 
    private Integer locationareacode; 
 
    private Integer baiduareaid; 
 
    private Double lat; 
 
    private Double lng; 
 
    private Integer areaid; 
    @Override 
    public void write(PreparedStatement statement) throws SQLException { 
        int index = 1;   
        statement.setString(index++, this.getMobilenetworkcode());   
        statement.setString(index++, this.getMobilecountrycode());   
        statement.setInt(index++, this.getCellid());   
        statement.setInt(index++, this.getLocationareacode()); 
        statement.setInt(index++, this.getBaiduareaid()); 
        statement.setDouble(index++, this.getLat()); 
        statement.setDouble(index++, this.getLng()); 
        statement.setInt(index, this.getAreaid()); 
    } 
 
    @Override 
    public void readFields(ResultSet resultSet) throws SQLException { 
         this.mobilenetworkcode = resultSet.getString(1); 
         this.mobilecountrycode = resultSet.getString(2); 
         this.cellid = resultSet.getInt(3); 
         this.locationareacode = resultSet.getInt(4); 
         this.baiduareaid = resultSet.getInt(5); 
         this.lat = resultSet.getDouble(6); 
         this.lng = resultSet.getDouble(7); 
         this.areaid = resultSet.getInt(8); 
    } 
 
    @Override 
    public void write(DataOutput out) throws IOException { 
        // TODO Auto-generated method stub 
         
    } 
 
    @Override 
    public void readFields(DataInput in) throws IOException { 
         
         
    } 
 
    public String getMobilenetworkcode() { 
        return mobilenetworkcode; 
    } 
 
    public void setMobilenetworkcode(String mobilenetworkcode) { 
        this.mobilenetworkcode = mobilenetworkcode; 
    } 
 
    public String getMobilecountrycode() { 
        return mobilecountrycode; 
    } 
 
    public void setMobilecountrycode(String mobilecountrycode) { 
        this.mobilecountrycode = mobilecountrycode; 
    } 
 
    public Integer getCellid() { 
        return cellid; 
    } 
 
    public void setCellid(Integer cellid) { 
        this.cellid = cellid; 
    } 
 
    public Integer getLocationareacode() { 
        return locationareacode; 
    } 
 
    public void setLocationareacode(Integer locationareacode) { 
        this.locationareacode = locationareacode; 
    } 
 
    public Integer getBaiduareaid() { 
        return baiduareaid; 
    } 
 
    public void setBaiduareaid(Integer baiduareaid) { 
        this.baiduareaid = baiduareaid; 
    } 
 
    public Double getLat() { 
        return lat; 
    } 
 
    public void setLat(Double lat) { 
        this.lat = lat; 
    } 
 
    public Double getLng() { 
        return lng; 
    } 
 
    public void setLng(Double lng) { 
        this.lng = lng; 
    } 
 
    public Integer getAreaid() { 
        return areaid; 
    } 
 
    public void setAreaid(Integer areaid) { 
        this.areaid = areaid; 
    } 
 
} 

转载于:https://my.oschina.net/u/2000675/blog/654627