如何使用hadoop.mapreduce.lib.output.MultipleOutputs使用oozie工作流创建目录结构？

Question

I am running my MR jobs using workflow:0.5". My use case to create key based directory structure for output. This is my configuration file:-

`           
        <configuration>
                <!-- These are important. -->
                <property>
                    <name>mapred.mapper.new-api</name>
                    <value>true</value>
                </property>
                <property>
                    <name>mapred.reducer.new-api</name>
                    <value>true</value>
                </property>
                <property>
                    <name>mapred.job.queue.name</name>
                    <value>${queue.name}
                    </value>
                </property>
                <property>
                    <name>mapreduce.map.class</name>
                    <value>com.a.b.c.Amapper</value>
                </property>
                <property>
                    <name>mapreduce.reduce.class</name>
                    <value>com.a.b.c.Areducer</value>
                </property>
                <property>
                    <name>mapred.output.key.class</name>
                    <value>org.apache.hadoop.io.Text</value>
                </property>
                <property>
                    <name>mapred.output.value.class</name>
                    <value>org.apache.hadoop.io.Text</value>
                </property>
                <property>
                    <name>mapreduce.outputformat.class</name>
                    <value>org.apache.hadoop.mapreduce.lib.output.MultipleOutputs
                    </value>
                </property>
                <property>
                    <name>mapred.input.dir</name>
                    <value>${inputDir}</value>
                </property>
                <property>
                    <name>mapred.output.dir</name>
                    <value>${outputDir}</value>
                </property>
            </configuration>`

In reducer, i want to create formatted directory structure using this code-

`public class Areducer extends Reducer<Text, Text, Text, Text> {
    private Text aggregatorRecord = new Text();
    private MultipleOutputs<Text, Text> out;

    public void setup(Context context) {
        out = new MultipleOutputs<Text, Text>(context);
    }

    public void reduce(Text aggregatorRecordKey,
            Iterable<Text> values, Context context)
            throws IOException, InterruptedException {
        /** 
           some business logic to do aggregation to set aggregatorRecord.
        */
        String plist = "Surname|Forename";
        Text t = new Text(plist);
        out.write(aggregatorRecordKey, aggregatorRecord, generateFileName(t));
    }

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

    private String generateFileName(Text k) {
        String[] kStr = k.toString().split("\\|");

        String sName = kStr[0];
        String fName = kStr[1];

        // example for k = Smith|John
        // output written to /user/hadoop/path/to/output/Smith/John-r-00000
        // (etc)
        return sName + "/" + fName;
    }

`

oozie workflow gives this exception

java.lang.NoSuchMethodException: org.apache.hadoop.mapreduce.lib.output.MultipleOutputs.<\\init>()

Can someone suggest the correct approach to use MultipleOutputs to create directory structure using oozie workflow?

Answer 1

Your problem is that MultipleOutputs is not an OutputFormat and so you don't set it as the output format for your job. I usually use a java class to configure and submit my MultipleOutputs jobs, but looking at your code I think what you need to to is set your output format type to TextOutputFormat and leave your references to reducer variable as they are.