saprk的udtf函数一列转多列应用

需求：查分outresult字段生成多个列

代码：

1.udtf部分

import com.alibaba.fastjson.JSONException;
import org.apache.commons.lang3.StringUtils;
import org.apache.hadoop.hive.ql.exec.UDFArgumentException;
import org.apache.hadoop.hive.ql.exec.UDFArgumentLengthException;
import org.apache.hadoop.hive.ql.metadata.HiveException;
import org.apache.hadoop.hive.ql.udf.generic.GenericUDTF;
import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory;
import org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory;

import java.util.ArrayList;

/**
 * 根据 outresult  进行解析
 * 继承org.apache.hadoop.hive.ql.udf.generic.GenericUDTF,
 * 实现initialize, process, close三个方法。
 */
public class ExplodeUDTF extends GenericUDTF {
	//用于规定列名
    private ArrayList<String> colName = new ArrayList();
    //用于指定列类型
    private ArrayList<ObjectInspector> resType = new ArrayList<>();

    @Override
    public StructObjectInspector initialize(ObjectInspector[] argOIs) throws UDFArgumentException {
        if (argOIs.length != 1) {
            throw new UDFArgumentLengthException(
                    "有且只能传入一个参数");
        }
        if (argOIs[0].getCategory() != ObjectInspector.Category.PRIMITIVE) {
            throw new UDFArgumentException(
                    "参数类型不匹配");
        }

        //新增字段的列名
       add_colName("col_1");
       add_colName("col_2");
       add_colName("col_3");
       add_colName("col_4");

        //新增字段的类型
  	   resType.add(PrimitiveObjectInspectorFactory.javaStringObjectInspector);
  	   resType.add(PrimitiveObjectInspectorFactory.javaStringObjectInspector);
  	   resType.add(PrimitiveObjectInspectorFactory.javaStringObjectInspector);
  	   resType.add(PrimitiveObjectInspectorFactory.javaStringObjectInspector);
  	   
        //定义返回的列名和列类型
        return ObjectInspectorFactory.getStandardStructObjectInspector(colName,resType);
    }
	// 此方法即为输入字段的处理逻辑
    @Override
    public void process(Object[] obj) throws HiveException {
        String input = obj[0].toString();
        if (StringUtils.isBlank(input)){
            return;
        }else try{

            //创建JSONObject数组，用于接收拿到的json对象
            Object[] res = new Object[4];

			//因业务需求，不多赘述
			res[0] = ''
			res[1] = ''   
            //写出结果
            forward(res);
        }catch (JSONException e){
            e.printStackTrace();
        }
    }

	//对需要清理的方法进行清理,未进行任何操作
    @Override
    public void close() throws HiveException {

    }

}

2.调用部分

public class ExplodeTest {
    public static void main(String[] args) {
        SparkSession spark = SparkSession.builder()
                //.master("local[*]")
                .appName("OutResultTest")
                .config("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
                .config("spark.sql.parquet.writeLegacyFormat", true)
                .enableHiveSupport()
                .getOrCreate();

        //读取sql文件
        InputStreamReader isr = new InputStreamReader(Run2Pid.class.getResourceAsStream("/pid2.sql"), StandardCharsets.UTF_8);
        String[] pidsql =new BufferedReader(isr).lines().collect(Collectors.joining("\n")).split(";");


        Dataset<Row> test = spark.sql(pidsql[1]);
        test.registerTempTable("test");

        //使用sql的方式注册udtf
        spark.sql("CREATE TEMPORARY FUNCTION myudtf as 'com.aaa.udf.ExplodeUDTF'");

        Dataset<Row> test2 = spark.sql(pidsql[2]);
        test2.registerTempTable("test2");

        spark.sql(pidsql[3]);


    }
}

若后续想列转行，可以和lateral view 联用，实现列转行。
可参考：
Lateral View语法应用