Hadoop MR编程

Hadoop开发job需要定一个Map/Reduce/Job（启动MR job，并传入参数信息），以下代码示例实现的功能：

1）将一个用逗号分割的文件，替换为“|”分割的文件；

2）对小文件合并，将文件合并为reduceNum个文件。

DataMap.java

package com.dx.fpd_load;

import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Mapper;

import java.io.IOException;

public class DataMap extends Mapper<LongWritable, Text, Text, Text> {
    private final Text key = new Text();

    @Override
    protected void map(LongWritable longWritable, Text value, Context context) throws IOException, InterruptedException {
        // 如果数据为空，则不进行处理，跳出map输入
        if (value.getLength() == 0) {
            return;
        }

        String newValue = value.toString().replace(",", "|") + "|NULL|NULL";
        String[] newValues = newValue.split("\\|");

        // 输入的文件路径
        String filePath = context.getInputSplit().toString().toUpperCase();

        // 如果路径包含了fpd_bak才进行处理否则不处理
        if (filePath.contains("fpd_bak".toUpperCase()) && newValues.length > 10) {
            key.set(newValues[6]); //objid

            context.write(key, new Text(newValue));
        }
    }
}

DataReducer.java

package com.dx.fpd_load;

import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.output.MultipleOutputs;

import java.io.IOException;

public class DataReducer extends Reducer<Text, Text, NullWritable, Text> {
    public MultipleOutputs multipleOutputs;
    public final Text key = new Text();

    @Override
    protected void setup(Context context) throws IOException, InterruptedException {
        multipleOutputs = new MultipleOutputs(context);
    }

    @Override
    protected void reduce(Text key, Iterable<Text> values, Context context) throws IOException, InterruptedException {
        for (Text text : values) {
            String data = text.toString();

            String[] p_days = context.getConfiguration().getStrings("p_day");
            String[] p_cities = context.getConfiguration().getStrings("p_city");

            String p_day = "p_day";
            if (p_days != null) {
                p_day = p_days[0];
            }
            String p_city = "p_city";
            if (p_cities != null) {
                p_city = p_cities[0];
            }

            multipleOutputs.write("fpdload", NullWritable.get(), new Text(data), "/thetenet/my_hive_db/fpd_new/p_day=" + p_day + "/p_city=" + p_city + "/fpd_data");
        }
    }

    @Override
    protected void cleanup(Context context) throws IOException, InterruptedException {
        multipleOutputs.close();
    }
}

DataJob.java

package com.dx.fpd_load;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.mapreduce.lib.output.MultipleOutputs;
import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;
import org.apache.hadoop.util.GenericOptionsParser;

public class FingerLib_Load_DataJob {
    public static void main(String[] args) throws Exception {
        Configuration conf = new Configuration();
        String[] otherArgs = new GenericOptionsParser(conf, args).getRemainingArgs();
        String p_city = otherArgs[0];
        String p_day = otherArgs[1];
        String reducerNum = otherArgs[2];
        String inputPath = otherArgs[3];
        String outputPath = otherArgs[4];

        if (p_day == null) {
            throw new Exception("p_day is null");
        }
        conf.set("p_day", p_day);
        if (p_city == null) {
            throw new Exception("p_city is null");
        }
        conf.set("p_city", p_city);

        Job job = Job.getInstance(conf);
        job.setJobName("LoadDataIntoFPD_p_city" + p_city + "_p_day_" + p_day);
        job.setJarByClass(DataJob.class);
        job.setMapperClass(DataMap.class);
        job.setReducerClass(DataReducer.class);
        job.setMapOutputKeyClass(Text.class);
        job.setMapOutputValueClass(Text.class);
        job.setOutputKeyClass(NullWritable.class);
        job.setOutputValueClass(Text.class);
        job.setNumReduceTasks(Integer.parseInt(reducerNum));

        MultipleOutputs.addNamedOutput(job, "fpdload", TextOutputFormat.class, NullWritable.class, Text.class);

        FileInputFormat.addInputPath(job, new Path(inputPath));
        FileOutputFormat.setOutputPath(job, new Path(outputPath));

        System.exit(job.waitForCompletion(true) ? 0 : 1);
    }
}

调用脚本：

#!/usr/bin/env bash
source /app/mylinux/login.sh
#./submit_fpdload.sh
DAY=$
CITY=$
REDUCER_NUMBER=$

JAR="/app/mylinux/service/dx-1.0-SNAPSHOT.jar"

MAIN_CLASS="com.dx.fpd_load.DataJob"
INPUT_PATH="/thetenet/my_hive_db/fpd_bak/p_day=$DAY/p_city=$CITY/"
OUT_DIR="/thetenet/my_hive_db/fpd_load_out/"

hadoop fs -rm -r /thetenet/my_hive_db/fpd_new/p_day=$DAY/p_city=$CITY/
hadoop fs -rm -r $OUT_DIR

time yarn jar $JAR $MAIN_CLASS $CITY $DAY $REDUCER_NUMBER $INPUT_PATH $OUT_DIR

#beeline -e "
#alter table my_hive_db.fpd_new add if not exists partition(p_day=$DAY,p_city=$CITY)
#location '/thetenet/my_hive_db/fpd_new/p_day=$DAY/p_city=$CITY/';"

echo "Complete..."