Hadoop综合项目——二手房统计分析（MapReduce篇）

文章目录

• Hadoop综合项目——二手房统计分析（MapReduce篇）
• • 0、 写在前面
  • 1、MapReduce统计分析
  • • 1.1 统计四大一线城市房价的最值
    • 1.2 按照城市分区统计二手房数量
    • 1.3 根据二手房信息发布时间排序统计
    • 1.4 统计二手房四大一线城市总价Top5
    • 1.5 基于二手房总价实现自定义分区全排序
    • 1.6 基于建造年份和房子总价的二次排序
    • 1.7 自定义类统计二手房地理位置对应数量
    • 1.8 统计二手房标签的各类比例
  • 2、数据及源代码
  • 3、总结

0、 写在前面

• Windows版本：Windows10
• Linux版本：Ubuntu Kylin 16.04
• JDK版本：Java8
• Hadoop版本：Hadoop-2.7.1
• Hive版本：Hive1.2.2
• IDE：IDEA 2020.2.3
• IDE：Pycharm 2021.1.3
• IDE：Eclipse3.8

1、MapReduce统计分析

通过MapReduce对最值、排序、TopN、自定义分区排序、二次排序、自定义类、占比等8个方面的统计分析

1.1 统计四大一线城市房价的最值

• 分析目的：

二手房房价的最值是体现一个城市经济的重要因素，也是顾客购买的衡量因素之一。

• 代码：

Driver端：

public class MaxMinTotalPriceByCityDriver {public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException {Configuration conf = new Configuration();Job job = Job.getInstance(conf, "MaxMinTotalPriceByCity");job.setJarByClass(MaxMinTotalPriceByCityDriver.class);job.setMapperClass(MaxMinTotalPriceByCityMapper.class);job.setReducerClass(MaxMinTotalPriceByCityReducer.class);job.setMapOutputKeyClass(Text.class);job.setMapOutputValueClass(IntWritable.class);job.setOutputKeyClass(Text.class);job.setOutputValueClass(Text.class);FileInputFormat.setInputPaths(job, new Path("datas/tb_house.txt"));FileOutputFormat.setOutputPath(job, new Path("MapReduce/out/MaxMinTotalPriceByCity"));job.waitForCompletion(true);}
}

• Mapper端：

public class MaxMinTotalPriceByCityMapper extends Mapper {private Text outk = new Text();private IntWritable outv = new IntWritable();@Overrideprotected void map(Object key, Text value, Context out) throws IOException, InterruptedException {String line = value.toString();String[] data = line.split("\t");outk.set(data[1]);      // cityoutv.set(Integer.parseInt(data[6]));        // totalout.write(outk, outv);}
}

Reducer端：

public class MaxMinTotalPriceByCityReducer extends Reducer {@Overrideprotected void reduce(Text key, Iterable values, Context context) throws IOException, InterruptedException {List totalList = new ArrayList();Iterator iterator = values.iterator();while (iterator.hasNext()) {totalList.add(iterator.next().get());}Collections.sort(totalList);int max = totalList.get(totalList.size() - 1);int min = totalList.get(0);Text outv = new Text();outv.set("房子总价最大、小值分别为：" + String.valueOf(max) + "万元," + String.valueOf(min) + "万元");context.write(key, outv);}
}

• 运行情况：

  ![tp](https://img-blog.csdnimg.cn/98168d46e7ac49e5adcd76a309ee6041.png)
  

• 结果：

  

1.2 按照城市分区统计二手房数量

• 分析目的：

二手房的数量是了解房子基本情况的维度之一，数量的多少在一定程度上体现了房子的受欢迎度。

• 代码：

Driver端：

public class HouseCntByCityDriver {public static void main(String[] args) throws Exception {args = new String[] { "/input/datas/tb_house.txt", "/output/HouseCntByCity" };Configuration conf = new Configuration();conf.set("fs.defaultFS", "hdfs://node01:9000");Job job = Job.getInstance(conf, "HouseCntByCity");job.setJarByClass(HouseCntByCityDriver.class);job.setMapperClass(HouseCntByCityMapper.class);job.setReducerClass(HouseCntByCityReducer.class);job.setMapOutputKeyClass(Text.class);job.setMapOutputValueClass(IntWritable.class);job.setOutputKeyClass(Text.class);job.setOutputValueClass(IntWritable.class);job.setPartitionerClass(CityPartitioner.class);job.setNumReduceTasks(4);FileInputFormat.addInputPath(job, new Path(args[0]));FileOutputFormat.setOutputPath(job, new Path(args[1]));job.waitForCompletion(true);}
}

Mapper端：

public class HouseCntByCityMapper extends Mapper {private Text outk = new Text();private IntWritable outv = new IntWritable(1);@Overrideprotected void map(Object key, Text value, Context context) throws IOException, InterruptedException {String line = value.toString();String[] data = line.split("\t");outk.set(new Text(data[1]));context.write(outk, outv);}
}

Reducer端：

public class HouseCntByCityReducer extends Reducer {@Overrideprotected void reduce(Text key, Iterable values, Context context) throws IOException, InterruptedException {int sum = 0;for (IntWritable val : values) sum += val.get();context.write(key, new IntWritable(sum));}
}

• 运行情况：

• 结果：

1.3 根据二手房信息发布时间排序统计

• 分析目的：

二手房的信息发布时间是了解房子基本情况的维度之一，在一定程度上，顾客倾向于最新的房源信息。

• 代码：

Driver端：

public class AcessHousePubTimeSortDriver {public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException {Configuration conf = new Configuration()Job job = Job.getInstance(conf, "AcessHousePubTimeSort");job.setJarByClass(AcessHousePubTimeSortDriver.class);job.setMapperClass(AcessHousePubTimeSortMapper.class);job.setReducerClass(AcessHousePubTimeSortReducer.class);job.setMapOutputKeyClass(Text.class);job.setMapOutputValueClass(IntWritable.class);job.setOutputKeyClass(Text.class);job.setOutputValueClass(IntWritable.class);FileInputFormat.setInputPaths(job, new Path("datas/tb_house.txt"));FileOutputFormat.setOutputPath(job, new Path("MapReduce/out/AcessHousePubTimeSort"));job.waitForCompletion(true);}
}

Mapper端：

public class AcessHousePubTimeSortMapper extends Mapper {private Text outk = new Text();private IntWritable outv = new IntWritable(1);@Overrideprotected void map(Object key, Text value, Context context) throws IOException, InterruptedException {String lines = value.toString();String data[] = lines.split("\t");String crawler_time = data[9], followInfo = data[4];String ct = crawler_time.substring(0, 10);int idx1 = followInfo.indexOf("|"), idx2 = followInfo.indexOf("发");String timeStr = followInfo.substring(idx1 + 1, idx2);String pubDate = "";try {pubDate = getPubDate(ct, timeStr);} catch (ParseException e) {e.printStackTrace();}outk.set(new Text(pubDate));context.write(outk, outv);}public String getPubDate(String ct, String timeStr) throws ParseException{SimpleDateFormat sdf = new SimpleDateFormat("yyyy-MM-dd");Date getTime = sdf.parse(ct);String getDate = sdf.format(getTime);Calendar calendar = Calendar.getInstance();calendar.setTime(getTime);if (timeStr.equals("今天")) {calendar.add(Calendar.DAY_OF_WEEK,-0);} else if (timeStr.contains("天")) {int i = 0;while (Character.isDigit(timeStr.charAt(i))) i++;int size = Integer.parseInt(timeStr.substring(0, i));calendar.add(Calendar.DAY_OF_WEEK, -size);} else {int i = 0;while (Character.isDigit(timeStr.charAt(i))) i++; int size = Integer.parseInt(timeStr.substring(0, i));calendar.add(Calendar.MONTH, -size);}Date pubTime = calendar.getTime();String pubDate = sdf.format(pubTime);return pubDate;}
}

Reducer端：

public class AcessHousePubTimeSortReducer extends Reducer {@Overrideprotected void reduce(Text key, Iterable values, Context context) throws IOException, InterruptedException {int sum = 0;for (IntWritable val : values) sum += val.get();context.write(key, new IntWritable(sum));}
}

• 运行情况：

• 结果：

1.4 统计二手房四大一线城市总价Top5

• 分析目的：

TopN是MapReduce分析最常见且必不可少的一个例子。

• 代码：

Driver端：

public class TotalPriceTop5ByCityDriver {public static void main(String[] args) throws Exception {args = new String[] {  "datas/tb_house.txt", "MapReduce/out/TotalPriceTop5ByCity" };Configuration conf = new Configuration();if (args.length != 2) {System.err.println("Usage: TotalPriceTop5ByCity  ");System.exit(2);}Job job = Job.getInstance(conf);job.setJarByClass(TotalPriceTop5ByCityDriver.class);job.setMapperClass(TotalPriceTop5ByCityMapper.class);job.setReducerClass(TotalPriceTop5ByCityReducer.class);job.setMapOutputKeyClass(Text.class);job.setMapOutputValueClass(IntWritable.class);job.setNumReduceTasks(1);job.setOutputKeyClass(Text.class);job.setOutputValueClass(Text.class);FileInputFormat.addInputPath(job, new Path(args[0]));FileOutputFormat.setOutputPath(job, new Path(args[1]));System.exit(job.waitForCompletion(true) ? 0 : 1);}
}

Mapper端：

public class TotalPriceTop5ByCityMapper extends Mapper {private int cnt = 1;private Text outk = new Text();private IntWritable outv = new IntWritable();@Overridepublic void map(Object key, Text value, Context context) throws IOException, InterruptedException {String line = value.toString();String[] data = line.split("\t");String city = data[1], totalPrice = data[6];outk.set(data[1]);outv.set(Integer.parseInt(data[6]));context.write(outk, outv);}
}

Reducer端：

public class TotalPriceTop5ByCityReducer extends Reducer {private Text outv = new Text();private int len = 0;@Overrideprotected void reduce(Text key, Iterable values, Context context) throws IOException, InterruptedException {List totalPriceList = new ArrayList();Iterator iterator = values.iterator();while (iterator.hasNext()) {totalPriceList.add(iterator.next().get());}Collections.sort(totalPriceList);int size = totalPriceList.size();String top5Str = "二手房总价Top5：";for (int i = 1; i <= 5; i++) {if (i == 5) {top5Str += totalPriceList.get(size - i) + "万元";} else {top5Str += totalPriceList.get(size - i) + "万元, ";}}outv.set(String.valueOf(top5Str));context.write(key, outv);}
}

• 运行情况：

• 结果：

1.5 基于二手房总价实现自定义分区全排序

• 分析目的：

自定义分区全排序可以实现不同于以往的排序方式，展示效果与默认全排序可以体现出一定的差别。

• 代码：

public class TotalOrderingPartition extends Configured implements Tool {static class SimpleMapper extends Mapper {@Overrideprotected void map(Object key, Text value, Context context) throws IOException, InterruptedException {IntWritable intWritable = new IntWritable(Integer.parseInt(key.toString()));context.write((Text) key, intWritable);}}static class SimpleReducer extends Reducer {@Overrideprotected void reduce(Text key, Iterable values, Context context) throws IOException, InterruptedException {for (IntWritable value : values) {context.write(value, NullWritable.get());}}}@Overridepublic int run(String[] args) throws Exception {Configuration conf = getConf();Job job = Job.getInstance(conf, "Total Order Sorting");job.setJarByClass(TotalOrderingPartition.class);job.setInputFormatClass(KeyValueTextInputFormat.class);FileInputFormat.addInputPath(job, new Path(args[0]));FileOutputFormat.setOutputPath(job, new Path(args[1]));job.setNumReduceTasks(3);job.setMapOutputKeyClass(Text.class);job.setMapOutputValueClass(IntWritable.class);job.setOutputKeyClass(IntWritable.class);job.setOutputValueClass(NullWritable.class);TotalOrderPartitioner.setPartitionFile(job.getConfiguration(), new Path(args[2]));InputSampler.Sampler sampler = new InputSampler.SplitSampler(5000, 10);InputSampler.writePartitionFile(job, sampler);job.setPartitionerClass(TotalOrderPartitioner.class);job.setMapperClass(SimpleMapper.class);job.setReducerClass(SimpleReducer.class);job.setJobName("TotalOrderingPartition");return job.waitForCompletion(true) ? 0 : 1;}public static void main(String[] args) throws Exception {args = new String[] { "datas/tb_house.txt", "MapReduce/out/TotalOrderingPartition/outPartition1", "MapReduce/out/TotalOrderingPartition/outPartition2" };int exitCode = ToolRunner.run(new TotalOrderingPartition(), args);System.exit(exitCode);}
}

• 运行情况：

• 结果：

…

…

…

1.6 基于建造年份和房子总价的二次排序

• 分析目的：

某些时候按照一个字段的排序方式并不能让我们满意，二次排则是解决这个问题的一个方法。

• 代码：

Driver端：

Mapper端：

Reducer端：

• 运行情况：

• 结果：

1.7 自定义类统计二手房地理位置对应数量

• 分析目的：

某些字段通过MapReduce不可以直接统计得到，这时采用自定义类的方式便可以做到。

• 代码：

自定义类：

public class HouseCntByPositionTopListBean implements Writable {private Text info;private IntWritable cnt;public Text getInfo() {return info;}public void setInfo(Text info) {this.info = info;}public IntWritable getCnt() {return cnt;}public void setCnt(IntWritable cnt) {this.cnt = cnt;}@Overridepublic void readFields(DataInput in) throws IOException {this.cnt = new IntWritable(in.readInt());}@Overridepublic void write(DataOutput out) throws IOException {out.writeInt(cnt.get());}@Overridepublic String toString() {String infoStr = info.toString();int idx = infoStr.indexOf("-");String city = infoStr.substring(0, idx);String position = infoStr.substring(idx + 1);return city + "#" + "[" + position + "]" + "#" + cnt;}
}

Driver端：

Mapper端：

Reducer端：

• 运行情况：

• 结果：

1.8 统计二手房标签的各类比例

• 分析目的：

占比分析同样是MapReduce统计分析的一大常用方式。

• 代码：

Driver端：

public class TagRatioByCityDriver {public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException {args = new String[] {"datas/tb_house.txt", "MapReduce/out/TagRatioByCity" };Configuration conf = new Configuration();Job job = Job.getInstance(conf);job.setJarByClass(TagRatioByCityDriver.class);job.setMapOutputKeyClass(Text.class);job.setMapOutputValueClass(IntWritable.class);job.setOutputKeyClass(Text.class);job.setOutputValueClass(Text.class);job.setMapperClass(TagRatioByCityMapper.class);job.setReducerClass(TagRatioByCityReducer.class);FileInputFormat.addInputPath(job, new Path(args[0]));FileOutputFormat.setOutputPath(job, new Path(args[1]));job.waitForCompletion(true);}
}

Mapper端：

public class TagRatioByCityMapper extends Mapper {private Text outk = new Text();private IntWritable outv = new IntWritable(1);@Overrideprotected void map(Object key, Text value, Context context) throws IOException, InterruptedException {String line = value.toString();String[] data = line.split("\t");String city = data[1], tag = data[8];if ("".equals(tag))  tag = "未知标签";outk.set(city + "-" + tag);context.write(outk, outv);}
}

Reducer端：

public class TagRatioByCityReducer extends Reducer {private Text outv = new Text();private int sum = 0;@Overrideprotected void reduce(Text key, Iterable values, Context context) throws IOException, InterruptedException {DecimalFormat df = new DecimalFormat("0.00");int cnt = 0;for (IntWritable value : values) {cnt += value.get();}String s = key.toString();String format = "";if (s.contains("上海")) {sum = 2995;format = df.format((double) cnt / sum * 100) + "%";} else if (s.contains("北京")) {sum = 2972;format = df.format((double) cnt / sum * 100) + "%";} else if (s.contains("广州")) {sum = 2699;format = df.format((double) cnt / sum * 100) + "%";} else {sum = 2982;format = df.format((double) cnt / sum * 100) + "%";}outv.set(format);context.write(key, outv);}
}

• 运行情况：

• 结果：

tp

2、数据及源代码

• Github

• Gitee

3、总结

MapReduce统计分析过程需要比较细心，「根据二手房信息发布时间排序统计」这个涉及到Java中日期类SimpleDateFormat和Date的使用，需要慢慢调试得出结果；统计最值和占比的难度并不高，主要在于统计要计算的类别的数量和总数量，最后二者相处即可；二次排序和自定义类难度较高，但一步一步来还是可以实现的。

结束！