阿里云MaxCompute自定义UDAF函数

环境：阿里云MaxCompute
阿里云MaxCompute的UDAF函数和hive的UDAF函数还是有所差别的。
需求：将拼接好的字段再次拼接，并将拼接好的结果去重，比如：北京，上海，天津；江西，湖北，湖南，北京；这两个字符串重新拼接后的结果应该为：北京，上海，天津，江西，湖北，湖南
这里我是用到了阿里云的MaxCompute Studio插件，关于该插件的下载和使用以及函数的发布大家可以参考官网https://helpcdn.aliyun.com/document_detail/27811.html?spm=a2c4g.11186623.6.601.64131f854SbjQa

import com.aliyun.odps.io.Text;
import com.aliyun.odps.io.Writable;
import com.aliyun.odps.udf.ExecutionContext;
import com.aliyun.odps.udf.UDFException;
import com.aliyun.odps.udf.Aggregator;
import com.aliyun.odps.udf.annotation.Resolve;


import java.util.*;

// TODO define input and output types, e.g. "double->double".
/**
 *
 * @author Liam
 * @description 将hscode的多个字符串重新聚合去重拼接
 * @date 2021-02-26 17:05:08
 **/

@Resolve({"string->string"})
public class HscodeStrMerge extends Aggregator {

    private Text result = new Text();


    @Override
    public void setup(ExecutionContext ctx) throws UDFException {

    }

    /**
     * 创建聚合Buffer
     *
     * @return Writable聚合buffer
     */
    @Override
    public Writable newBuffer() {
        return new Text();
    }

    /**
     * @param buffer 聚合buffer，区内合并
     * @param args   SQL中调用UDAF时指定的参数，不能为null，但是args里面的元素可以为null，代表对应的输入数据是null
     * @throws UDFException
     */
    @Override
    public void iterate(Writable buffer, Writable[] args) throws UDFException {
        // 区内合并
        Text buf = (Text) buffer;
        for (Writable arg : args) {
            if (!Objects.isNull(arg) && !"".equals(arg.toString())){
                buf.set(arg + "," + buf);
            }
        }
    }

    /**
     * @param buffer  聚合buffer，区间合并
     * @param partial 分片聚合结果
     * @throws UDFException
     */
    @Override
    public void merge(Writable buffer, Writable partial) throws UDFException {
        // 区间合并
        Text buf = (Text) buffer;
        Text par = (Text) partial;
        if (!Objects.isNull(partial) && !"".equals(par.toString())){
            buf.set(par + "," + buf);
        }
    }

    /**
     * 生成最终结果
     *
     * @param buffer
     * @return Object UDAF的最终结果
     * @throws UDFException
     */
    @Override
    public Writable terminate(Writable buffer) throws UDFException {
        HashSet<String> set = new HashSet<>();
        StringBuilder strResult = new StringBuilder();
        // 生成最终结果
        Text buf = (Text) buffer;
        String[] split = buf.toString().split(",");
        set.addAll(Arrays.asList(split));
        for (String s : set) {
            strResult.append(s).append(",");
        }
        if (strResult.length()!=0){
             result.set(strResult.substring(0,strResult.length()-1));
            return result;
        }else {
            return null;
        }
    }

    @Override
    public void close() throws UDFException {

    }

}

更具体的实现可以参考：https://blog.csdn.net/beautiful_huang/article/details/107026903