wordcount.java

统计文件内不同文字出现的频率

package eclipse.workspace.Word.src;
import eclipse.workspace.Word.src.WordCount;

import java.io.FileNotFoundException;
import java.util.*;


public class WordCount {//implements java.lang.Comparable<Object>
	FileWordReader fwr;
	
	Map<String,Integer> wordList;
	int topNum; //the most popular words number,i,e top "10"
	int sum=0;	//count the topNum words appeared time
	Object[] obj; //to set the sorted wordlist
	
	private int[] value;
	private String[] key;
	
	public WordCount(String name) 
	{
		try {
			fwr=new FileWordReader(name);
			wordList=new HashMap<String,Integer>();
			this.setTopNum(10);
			this.setWordMap();
			this.sortTopWordValue();
			this.setLists();
			
			
		} catch (FileNotFoundException e) {
			e.printStackTrace();
		}
		
	} 
	
	public void setTopNum(int a)
	{
		topNum=a;
	}
	
	public void setWordMap() //put different words in the map and count their appeared time
	{
		int count;
		String key=fwr.getNextWord();
		while(key!=null)
		{
			count=1;
			key=key.toLowerCase();
			if(wordList.containsKey(key))
			{
				count=wordList.get(key);
				count++;
				wordList.put(key, count);
			}
			else
			{
				wordList.put(key, count);
			}
			key=fwr.getNextWord();
		}
	}
	
	public int getWordMapLength()
	{
		return wordList.size();
		
	}
	
	public int[] getValueList()
	{
		return value;
		
	}
	
	public void setLists()
	{
		key=new String[wordList.size()];
		value=new int[wordList.size()];
		for(int i=0;i<wordList.size();i++)
		{
			Map.Entry<String, Integer> entry=(Map.Entry<String, Integer>)obj[i];
			value[i]=entry.getValue();
			key[i]=entry.getKey();
		}
	}
	
	public String[] getWordList()
	{
		return key;
	}
	public void sortTopWordValue()//sort and print the most popular words in the file
	
	{
	
		List myList=new ArrayList(wordList.entrySet());
		Collections.sort(myList, new Comparator(){
			public int compare(Object o1,Object o2){
				Map.Entry< String, Integer> obj1=(Map.Entry<String, Integer>)o1;
				Map.Entry<String, Integer> obj2=(Map.Entry<String, Integer>)o2;
				return obj2.getValue().compareTo(obj1.getValue());
			}
		}
		);
		
		
		obj=myList.toArray();
		
		for(int i=1;i<=topNum;i++)
		{
			Map.Entry<String, Integer> entry=(Map.Entry<String, Integer>)obj[i-1];
			System.out.println("Top"+i+":\t"+entry.getKey()+" \tAppeared Time:"+entry.getValue());
			sum+=entry.getValue();
		}
		
		
		
		/*for(String k:wordList.keySet())
		{
			System.out.println("×Key:"+k+" Value:"+wordList.get(k));
		}*/
		}
	
	public Object[] getSortedList()
	{
		if(obj==null)
			this.sortTopWordValue();
		return obj;
	}

	
	public void grapPWord(long num,Object[] obj)
	{
		System.out.println("The graphic show of the top "+topNum+" Words(Percentage)");
		
		for(int i=0;i<this.wordList.size();i++)
		{
			Map.Entry<String, Integer> entry=(Map.Entry<String, Integer>)obj[i];
			System.out.printf("%20s",entry.getKey()+"\t");
			for(int j=0;j<=entry.getValue()*1000/num;j++)
				System.out.print('_');
			System.out.printf("%.2f", (float)entry.getValue()*100/num);
			System.out.println("%");
		}
		
	}
}