Hadoop 2.5.1學習筆記6:不一樣數據來源的聯結代碼範例

下面的程序有個bug,應該是

Path[] cacheFiles = context.getLocalCacheFiles();

if (null != cacheFiles && cacheFiles.length > 0) {
          readFile(cacheFiles[0].toString(), context);
}

代碼中未修正，讀者自行修正，謝謝！

數據源來源於2個： 1個是HDFS,數據量大，還一個是mongodb,數據量小。

須要2個來源一塊兒作數據聯結，而後分析。代碼範例以下：

-----------------------------------------------------------------------

主要思路：

從mongodb中導入數據大hdfs,而後經過hadoop的分發機制分發此文件

到全部計算節點上做爲「背景」數據。

而後計算節點的map類的setup函數中讀取此文件便可。

----------------------------------------------------------------------- 實際代碼以下：

package com.dew.task;

import java.io.BufferedReader;
import java.io.FileReader;
import java.io.IOException;
import java.net.URI;
import java.util.ArrayList;
import java.util.Hashtable;
import java.util.Set;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.conf.Configured;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.mapreduce.lib.output.NullOutputFormat;
import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;
import org.apache.hadoop.util.Tool;
import org.apache.hadoop.util.ToolRunner;

import com.mongodb.BasicDBObject;

public class ComputeProfileHDFS extends Configured implements Tool {

	// map
	public static class MapClass extends Mapper<LongWritable, Text, Text, Text> {

		private Hashtable<String, String> joinData = new Hashtable<String, String>();

		private void readFile(String file) {
			BufferedReader joinReader = null;
			String line = null;
			try {
				joinReader = new BufferedReader(new FileReader(file));
				while ((line = joinReader.readLine()) != null) {
					String[] array = line.split("\t");
					if (null == array || array.length < 2)
						continue;

					String pkg = array[0];
					if (null == pkg || pkg.length() <= 0)
						continue;

					String tagStr = array[1];
					if (null == tagStr)
						continue;
					tagStr = tagStr.trim();
					if (tagStr.length() <= 0)
						continue;
					joinData.put(pkg, tagStr);
					System.out.println("[map,setup] " + pkg + "  |  " + tagStr);
				}
			} catch (Exception e) {
				// XXX
				System.out
						.println("--------------------------------------------\n"
								+ e.toString());
			} finally {
				if (null != joinReader)
					try {
						joinReader.close();
					} catch (IOException e) {
						e.printStackTrace();
					}
			}

		}

		protected void setup(Context context) throws java.io.IOException,
				java.lang.InterruptedException {
			try {
				// Configuration conf = context.getConfiguration();
				URI[] cacheFiles = context.getCacheFiles();
				if (null != cacheFiles && cacheFiles.length > 0) {

					readFile(cacheFiles[0].getPath().toString());

				}
			} catch (IOException e) {
				// xxx
			}
		}

		public void map(LongWritable key, Text value, Context context)
				throws IOException, InterruptedException {

			// key neglected
			if (null == value)
				return;
			String content = value.toString();
			if (null == content || content.trim().length() == 0)
				return;
			// split
			String[] strArray = content.split("\t");
			if (null == strArray || strArray.length < 29)
				return;
			String sender = strArray[12].trim();
			String receiver = strArray[14].trim();
			String pkg = strArray[28].trim();
			if (null == sender || sender.length() == 0 || null == receiver
					|| receiver.length() == 0 || null == pkg
					|| pkg.length() == 0) {
				return;
			}
			String tags = this.joinData.get(pkg);
			if (null == tags || tags.trim().length() == 0)
				return;
			// okay,output it
			System.out.println("sender---" + sender + "   tags---" + tags);
			System.out.println("receiver---" + receiver + "   tags---" + tags);
			context.write(new Text(sender), new Text(tags));
			context.write(new Text(receiver), new Text(tags));
		}

	}

	public static class Combiner extends Reducer<Text, Text, Text, Text> {
		public void reduce(Text key, Iterable<Text> values, Context context)
				throws IOException, InterruptedException {
			String totalTags = "";
			for (Text tag : values) {
				totalTags += " " + tag;
			}
			totalTags = totalTags.trim();
			if (totalTags.length() <= 0)
				return;
			// okay
			// System.out.println("combiner function invoked....********************");
			context.write(key, new Text(totalTags));

		}

	}

	public static class Reduce extends Reducer<Text, Text, Text, Text> {
		public void reduce(Text key, Iterable<Text> values, Context context)
				throws IOException, InterruptedException {

			String totalTags = "";
			for (Text tag : values) {
				totalTags += " " + tag;
			}
			totalTags = totalTags.trim();
			if (totalTags.length() <= 0)
				return;
			// okay,let us do it now!
			String[] tagArray = totalTags.split(" ");
			if (null == tagArray || tagArray.length <= 0)
				return;
			// context.write(arg0, arg1);
			HttpApiClient.writeRecord(key.toString(), tagArray);
		}

		protected void cleanup(Context context) throws java.io.IOException,
				java.lang.InterruptedException {
			HttpApiClient.submit();
		}

	}

	public static class HttpApiClient {

		private static ArrayList<BasicDBObject> persons = new ArrayList<BasicDBObject>();
		private static int lock = 50;

		public static void writeRecord(String key, String[] arrays) {
			System.out.println("writeRecord 4");
			// 數據校驗
			if (null == key || key.length() <= 0 || null == arrays
					|| arrays.length <= 0)
				return;
			// 字符串數組--->數量統計
			Hashtable<String, Integer> table = new Hashtable<String, Integer>();
			for (String tag : arrays) {
				Integer number = table.get(tag);
				int count = (null == number ? 0 : number.intValue());
				count++;
				table.put(tag, count);
			}
			// 構造單我的標籤
			ArrayList<BasicDBObject> tagDocument = new ArrayList<BasicDBObject>();
			Set<String> tagSet = table.keySet();
			for (String tag : tagSet) {
				BasicDBObject doc = new BasicDBObject();
				doc.put("n", tag);
				doc.put("#", table.get(tag).intValue());
				tagDocument.add(doc);
			}
			// 構造單我的的文檔
			BasicDBObject person = new BasicDBObject();
			person.put("_id", key);
			person.put("t", tagDocument);
			System.out.println("*************************************");
			System.out.println(person.toString());
			System.out.println("*************************************");
			// 加入到全局當中
			persons.add(person);
			if (persons.size() >= lock) {
				submit();
			}
		}

		public static void submit() {
			// 提交上去併發送http請求...
			// persons.clear();
		}

	}

	@Override
	public int run(String[] args) throws Exception {

		Configuration conf = getConf();
		Job job = new Job(conf, "ComputeProfileHDFSPlusMongoDB");

		// add distributed file
		job.addCacheFile(new Path(args[1]).toUri());
		// DistributedCache.addCacheFile(new Path(args[1]).toUri(),
		// job.getConfiguration());

		// prepare
		FileInputFormat.setInputPaths(job, new Path(args[2]));
		FileOutputFormat.setOutputPath(job, new Path(args[3]));
		// FileOutputFormat.setOutputPath(job, new Path(args[2]));
		job.setJobName("ComputeProfileHDFSPlusMongoDB");
		job.setMapperClass(MapClass.class);
		job.setCombinerClass(Combiner.class);
		job.setReducerClass(Reduce.class);// job.setNumReduceTasks(0);
		job.setInputFormatClass(TextInputFormat.class);
		job.setOutputKeyClass(Text.class);
		job.setOutputValueClass(Text.class);
		job.setOutputFormatClass(NullOutputFormat.class);

		// execute
		int exitCode = job.waitForCompletion(true) ? 0 : 1;
		try {
			FileSystem fs = FileSystem.get(conf);
			fs.delete(new Path(args[1]));
			fs.delete(new Path(args[3]));
		} catch (Exception e) {

		}
		return exitCode;
	}

	public static void main(String[] args) throws Exception {

		int res;
		res = ToolRunner.run(new Configuration(), new PullMongoDB(), args);

		res = ToolRunner.run(new Configuration(), new ComputeProfileHDFS(),
				args);
		System.exit(res);
	}

}

 

package com.dew.task;

import java.util.ArrayList;
import java.util.List;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.conf.Configured;
import org.apache.hadoop.fs.FSDataOutputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.util.Tool;

import com.mongodb.BasicDBObject;
import com.mongodb.DB;
import com.mongodb.DBCollection;
import com.mongodb.DBCursor;
import com.mongodb.MongoClient;
import com.mongodb.ServerAddress;

public class PullMongoDB extends Configured implements Tool {

	@Override
	public int run(String[] args) throws Exception {
		if (null == args || args.length < 4) {
			return 0;
		}
		List list = new ArrayList();
		String[] array = args[0].split(":");
		list.add(new ServerAddress(array[0], Integer.parseInt(array[1])));
		MongoClient mongoClient = new MongoClient(list);
		DB database = mongoClient.getDB("" + array[2]);
		DBCollection collection = database.getCollection("" + array[3]);

		// 開始查詢
		BasicDBObject query = new BasicDBObject();
		query.put("pkg", new BasicDBObject("$exists", true));
		query.put("tags", new BasicDBObject("$exists", true));
		BasicDBObject fields = new BasicDBObject();
		fields.put("pkg", 1);
		fields.put("tags", 1);

		// 準備寫入的HDFS文件
		Configuration conf = new Configuration();
		FileSystem hdfs = FileSystem.get(conf);
		FSDataOutputStream outHandler = hdfs.create(new Path("" + args[1]));
		
		// 準備寫入
		DBCursor cursor = collection.find(query, fields);
		while (cursor.hasNext()) {
			BasicDBObject record = (BasicDBObject) cursor.next();
			String pkg = record.getString("pkg");
			ArrayList<String> als = (ArrayList<String>) record.get("tags");
			String tags = "";
			for (String s : als) {
				tags += " " + s.trim();
			}
			tags = tags.trim();
			String finalString = pkg + "\t" + tags + System.getProperty("line.separator");
			outHandler.write(finalString.getBytes("UTF8"));
		}
		
		// 移除句柄
		outHandler.close();
		cursor.close();
		mongoClient.close();
		return 0;
	}

}

 

測試經過

 

 補充： job.setJarByClass(ComputeProfileHDFS.class); ！！！