使用mapreduce在hadoop中进行文件比较

Question

A .txt

col1       col2    col3    col4    col5
A           120     140     160     180
B           200     220     240     260
D           400     420     440     460

B.txt

col1      col2    col3    col4    col5
A          110     140     160     180
B          200     220     240     260
C          600     620     640     660

输出-文件

A          120     140     160     180
A          110     140     160     180
B          200     220     240     260
D          400     420     440     460
C          600     620     640     660

1）col1和col2是这个中的主键，任何键都被改变，然后我们显示两个记录，如

in A.txt contain 1st Records:- A          120      140     160     180
in B.txt contain 1st Records:- A          110      140     160     180 

此col2中的

已更改，因此我必须显示两条记录

2）如果两个文件的记录没有变化（我的意思相同）我们只需要显示一条记录

3）在两个文件中显示所有其他记录

最终输出应该如下所示

输出-文件

A          120     140     160     180
A          110     140     160     180
B          200     220     240     260
D          400     420     440     460
C          600     620     640     660

Answer 1

使用PIG.Load这两个文件，合并记录然后将它们分开。

A = LOAD 'A.txt' USING PigStorage('\t');
B = LOAD 'B.txt' USING PigStorage('\t');
C = UNION A,B;
D = DISTINCT C;
DUMP D;

Answer 2

这是mapreduce解决方案：
将2个或更多文件放在一个目录中（输入 - arg1），它会将所有文件合并到一个满足您所有要求的文件。它还匹配col3到一个键的非macthing行的结束（col1 + col2）请参阅注释以获取更多信息......

public class FileCompare  extends Configured implements Tool{

    public static class FileComapreMapper extends Mapper<Object, Text, Text, Text> {
        int lineno=0;

        public void map(Object key, Text value, Context context) throws IOException, InterruptedException{
            try{
                lineno++;
                System.out.println(lineno + " -> " + value);
                //skip header - uncomment this line to include header in output
                if(lineno == 1) return; 

                String[] fields = value.toString().split("\\s+");//assuming input recs are whitespace seperated
                String col1_col2 = fields[0] + "," + fields[1]; //key
                String col3tolast="";
                for(int i=2; i < fields.length;i++)
                    col3tolast+=fields[i] + ","; //values

               col3tolast=col3tolast.substring(0, col3tolast.length()-1); //remove last char(',')
               context.write(new Text(col1_col2), new Text(col3tolast)); //send key, value pairs to reducer
            }catch(Exception e){
                System.err.println("Invaid Data at line: " + lineno + " Error: " + e.getMessage());
            }
        }   
    }

    public  static class FileComapreReducer extends Reducer<Text, Text, Text, Text> {
        @Override
        public void reduce(Text key, Iterable<Text> values, Context context) 
                throws IOException, InterruptedException {
            //Get unique col3 to last value
            Set<Text> uniqueCol3tolast = new HashSet<Text>();
            for(Text record : values)
                uniqueCol3tolast.add(record);
            //write key + value
            for(Text col3tolast:uniqueCol3tolast) //outputing tab delimited recs
                context.write(new Text(key.toString().replaceAll(",", "\t")), 
                        new Text(col3tolast.toString().replaceAll(",", "\t")));     
        }
    }

    public static void main(String[] args) throws Exception {
        int res = ToolRunner.run(new Configuration(), new FileCompare(), args);
        System.exit(res);
    }

    public int run(String[] args) throws Exception {
        if (args.length != 2) {
            System.err.println("Usage: <in> <out>");
            System.exit(2);
        }
        Configuration conf = this.getConf();
        Job job = Job.getInstance(conf, "merge-two-files");
        job.setJarByClass(FileCompare.class);
        job.setMapperClass(FileComapreMapper.class);
        job.setReducerClass(FileComapreReducer.class);
        job.setMapOutputKeyClass(Text.class);
        job.setMapOutputValueClass(Text.class);
        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(Text.class);

        FileInputFormat.addInputPath(job, new Path(args[0]));
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        FileSystem fs = null;
        Path dstFilePath = new Path(args[1]);
        try {
            fs = dstFilePath.getFileSystem(conf);
            if (fs.exists(dstFilePath))
                fs.delete(dstFilePath, true);
        } catch (IOException e1) {
            e1.printStackTrace();
        }
        return job.waitForCompletion(true) ? 0 : 1;
    } 
}