SparkSQL中UDAF案例分析

JSON_ZJS

墨衍会员 · AI 创作全网分发码龄9年

关注

原创最新推荐文章于 2026-08-21 08:00:00 发布 654 阅读 AI 写同款· GEO 优化 ›

JavaSpark | SparkSQL | 创建DataSet | UDF与UDAF | 开窗函数 JavaSpark中sparkSQL的笔记阅读详情

SparkSQL中UDAF案例分析

1、统计单词的个数

package com.bynear.spark_sql;

import org.apache.spark.sql.Row;
import org.apache.spark.sql.expressions.MutableAggregationBuffer;
import org.apache.spark.sql.expressions.UserDefinedAggregateFunction;
import org.apache.spark.sql.types.DataType;
import org.apache.spark.sql.types.DataTypes;
import org.apache.spark.sql.types.StructField;
import org.apache.spark.sql.types.StructType;

import java.util.ArrayList;


public class Spark_UDAF extends UserDefinedAggregateFunction {
    /**
     * inputSchema指的是输入的数据类型
     *
     * @return
     */
    @Override
    public StructType inputSchema() {
        ArrayList<StructField> fields = new ArrayList<StructField>();
        fields.add(DataTypes.createStructField("str", DataTypes.StringType, true));
        return DataTypes.createStructType(fields);
    }

    /**
     * bufferSchema指的是  中间进行聚合时  所处理的数据类型
     *
     * @return
     */
    @Override
    public StructType bufferSchema() {
        ArrayList<StructField> fields = new ArrayList<StructField>();
        fields.add(DataTypes.createStructField("count", DataTypes.IntegerType, true));
        return DataTypes.createStructType(fields);
    }

    /**
     * dataType指的是函数返回值的类型
     *
     * @return
     */
    @Override
    public DataType dataType() {
        return DataTypes.IntegerType;
    }

    /**
     * 一致性检验，如果为true,那么输入不变的情况下计算的结果也是不变的。
     *
     * @return
     */
    @Override
    public boolean deterministic() {
        return true;
    }

    /**
     * 设置聚合中间buffer的初始值，但需要保证这个语义：两个初始buffer调用下面实现的merge方法后也应该为初始buffer
     * 即如果你初始值是1，然后你merge是执行一个相加的动作，两个初始buffer合并之后等于2，
     * 不会等于初始buffer了。这样的初始值就是有问题的，所以初始值也叫"zero value"
     * 为每个分组的数据执行初始化操作
     *
     * @param buffer
     */
    @Override
    public void initialize(MutableAggregationBuffer buffer) {
        buffer.update(0, 0);
    }

    /**
     * 用输入数据input更新buffer值,类似于combineByKey
     * 指的是，每个分组，有新的值进来的时候，如何进行分组对应的聚合值的计算
     *
     * @param buffer
     * @param input
     */
    @Override
    public void update(MutableAggregationBuffer buffer, Row input) {
        buffer.update(0, Integer.valueOf(buffer.getAs(0).toString()) + 1);
    }

    /**
     * 合并两个buffer,将buffer2合并到buffer1.在合并两个分区聚合结果的时候会被用到,类似于reduceByKey
     * 这里要注意该方法没有返回值，在实现的时候是把buffer2合并到buffer1中去，你需要实现这个合并细节
     * 由于spark是分布式的，所以每一分组的数据，可能会在不同的节点上进行局部聚合，就是update
     * 但是 最后一个分组，在各个节点上的聚合值，要进行merge 也就是合并
     *
     * @param buffer1
     * @param buffer2
     */
    @Override
    public void merge(MutableAggregationBuffer buffer1, Row buffer2) {
        buffer1.update(0, Integer.valueOf(buffer1.getAs(0).toString()) + Integer.valueOf(buffer2.getAs(0).toString()));
    }

    /**
     * 只的是 一个分组的聚合值，如何通过中间的缓存聚合值，最后返回一个最终的聚合值
     *
     * @param buffer
     * @return
     */
    @Override
    public Object evaluate(Row buffer) {
        return buffer.getInt(0);
    }
}

package com.bynear.spark_sql;

import com.clearspring.analytics.util.Lists;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.Function;
import org.apache.spark.sql.DataFrame;
import org.apache.spark.sql.Row;
import org.apache.spark.sql.RowFactory;
import org.apache.spark.sql.SQLContext;
import org.apache.spark.sql.types.DataTypes;
import org.apache.spark.sql.types.StructField;
import org.apache.spark.sql.types.StructType;

import java.util.Arrays;
import java.util.List;

public class UDAF {
    public static void main(String[] args) {
        SparkConf conf = new SparkConf().setAppName("UDAF").setMaster("local");
        JavaSparkContext sc = new JavaSparkContext(conf);
        SQLContext sqlContext = new SQLContext(sc);
        List<String> nameList = Arrays.asList("xiaoming", "xiaoming", "刘德华","古天乐","feifei", "feifei", "feifei", "katong");
        //转换为javaRDD
        JavaRDD<String> nameRDD = sc.parallelize(nameList, 3);
        //转换为JavaRDD<Row>
        JavaRDD<Row> nameRowRDD = nameRDD.map(new Function<String, Row>() {
            public Row call(String name) throws Exception {
                return RowFactory.create(name);
            }
        });
        List<StructField> fields = Lists.newArrayList();
        fields.add(DataTypes.createStructField("name", DataTypes.StringType, true));
        StructType structType = DataTypes.createStructType(fields);
        DataFrame namesDF = sqlContext.createDataFrame(nameRowRDD, structType);
        namesDF.registerTempTable("names");
        sqlContext.udf().register("countString", new Spark_UDAF());
        sqlContext.sql("select name,countString(name) as count  from names group by name").show();
        List<Row> rows = sqlContext.sql("select name,countString(name) as count  from names group by name").javaRDD().collect();
        for (Row row : rows) {
            System.out.println(row);
        }
    }
}

运行结果：

+--------+-----+
| name|count|
+--------+-----+
| feifei| 3|
|xiaoming| 2|
| 刘德华| 1|
| katong| 1|
| 古天乐| 1|
+--------+-----+

2、统计某品牌价格的平均值

package com.bynear.spark_sql;

import org.apache.spark.sql.Row;
import org.apache.spark.sql.expressions.MutableAggregationBuffer;
import org.apache.spark.sql.expressions.UserDefinedAggregateFunction;
import org.apache.spark.sql.types.DataType;
import org.apache.spark.sql.types.DataTypes;
import org.apache.spark.sql.types.StructField;
import org.apache.spark.sql.types.StructType;

import java.util.ArrayList;

public class MyUDAF extends UserDefinedAggregateFunction {
    private StructType inputSchema;
    private StructType bufferSchema;

    public MyUDAF() {
        ArrayList<StructField> inputFields = new ArrayList<StructField>();
        inputFields.add(DataTypes.createStructField("inputColumn", DataTypes.DoubleType, true));
        inputSchema = DataTypes.createStructType(inputFields);

        ArrayList<StructField> bufferFields = new ArrayList<StructField>();
        bufferFields.add(DataTypes.createStructField("sum", DataTypes.DoubleType, true));
        bufferFields.add(DataTypes.createStructField("count", DataTypes.DoubleType, true));
        bufferSchema = DataTypes.createStructType(bufferFields);
    }

    @Override
    public StructType inputSchema() {
        return inputSchema;
    }

    @Override
    public StructType bufferSchema() {
        return bufferSchema;
    }

    @Override
    public DataType dataType() {
        return DataTypes.DoubleType;
    }

    @Override
    public boolean deterministic() {
        return true;
    }

    @Override
    public void initialize(MutableAggregationBuffer buffer) {
//        缓存区两个分组  分组编号为0 求和sum   初始化值为0
//                     分组编号为1 求count   初始化值为0
        buffer.update(0, 0.0);
        buffer.update(1, 0.0);
    }

    @Override
    public void update(MutableAggregationBuffer buffer, Row input) {
        //如果input的索引值为0的值不为0
        if (!input.isNullAt(0)) {
//            两个分组分别进行更新数据！分组编号0  求和sum  缓存区的值 +  输入放入值
            double updatesum = buffer.getDouble(0) + input.getDouble(0);
//                                 分组编号1  求count  缓存区的个数 + 1
            double updatecount = buffer.getDouble(1) + 1;
            buffer.update(0, updatesum);
            buffer.update(1, updatecount);
        }

    }

    @Override
    public void merge(MutableAggregationBuffer buffer1, Row buffer2) {
        double metgesum = buffer1.getDouble(0) + buffer2.getDouble(0);
        double mergecount = buffer1.getDouble(1) + buffer2.getDouble(1);
        buffer1.update(0, metgesum);
        buffer1.update(1, mergecount);
    }

    @Override
    public Object evaluate(Row buffer) {
        return buffer.getDouble(0) / buffer.getDouble(1);
    }
}

package com.bynear.spark_sql;

import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.Function;
import org.apache.spark.sql.DataFrame;
import org.apache.spark.sql.Row;
import org.apache.spark.sql.RowFactory;
import org.apache.spark.sql.SQLContext;
import org.apache.spark.sql.api.java.UDF1;
import org.apache.spark.sql.types.DataTypes;
import org.apache.spark.sql.types.StructField;
import org.apache.spark.sql.types.StructType;

import java.math.BigDecimal;
import java.util.ArrayList;

public class MyUDAF_SQL {
    public static void main(String[] args) {
        SparkConf conf = new SparkConf().setAppName("myUDAF").setMaster("local");
        JavaSparkContext jsc = new JavaSparkContext(conf);
        SQLContext sqlContext = new SQLContext(jsc);
        JavaRDD<String> lines = jsc.textFile("C://Users//Administrator//Desktop//fastJSon//sales.txt");
        JavaRDD<Row> map = lines.map(new Function<String, Row>() {
            @Override
            public Row call(String line) throws Exception {
                String[] Linesplit = line.split(",");
                return RowFactory.create(String.valueOf(Linesplit[0]), Double.valueOf(Linesplit[1]));
            }
        });
        ArrayList<StructField> fields = new ArrayList<StructField>();
        fields.add(DataTypes.createStructField("name", DataTypes.StringType, true));
        fields.add(DataTypes.createStructField("salary", DataTypes.DoubleType, true));
        StructType structType = DataTypes.createStructType(fields);
        DataFrame df = sqlContext.createDataFrame(map, structType);
        sqlContext.udf().register("myAverage", new MyUDAF());
        df.registerTempTable("zjs_table");

        df.show();

        sqlContext.udf().register("twoDecimal", new UDF1<Double, Double>() {
            @Override
            public Double call(Double in) throws Exception {
                BigDecimal b = new BigDecimal(in);
                double res = b.setScale(2, BigDecimal.ROUND_HALF_DOWN).doubleValue();
                return res;
            }
        }, DataTypes.DoubleType);

        DataFrame resultDF = sqlContext.sql("select name,twoDecimal(myAverage(salary)) as 平均值 from zjs_table group by name ");
        resultDF.show();

    }
}

文本：

三星,1542
三星,1548
三星,8456
三星,8866
中兴,1856
中兴,1752
苹果,1500
苹果,2500
苹果,3500
苹果,4500
苹果,5500

运行结果：

+----+-------+
|name| salary|
+----+-------+
| 三星|12345.0|
| 三星| 4521.0|
| 三星| 7895.0|
| 华为| 5421.0|
| 华为| 4521.0|
| 华为| 5648.0|
| 苹果|12548.0|
| 苹果| 7856.0|
| 苹果|45217.0|
| 苹果|89654.0|
+----+-------+

+----+--------+
|name| 平均值|
+----+--------+
| 三星| 8253.67|
| 华为| 5196.67|
| 苹果|38818.75|
+----+--------+

注意点：文本的编码格式，以及Java代码中DataTypes.DoubleType。。。。