// CSE 143, Winter 2011, Marty Stepp
// This program counts the number of words in a large file.
// It demonstrates the use of a Set collection.

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

// arraylist smallmoby 2.6s  mobydick 7.2s  bible 16.7s
// treeset   smallmoby 0.47s mobydick 0.63s bible 1.3s

public class WordCount {
    public static void main(String[] args) throws FileNotFoundException {
        Set<String> words = new TreeSet<String>();
        
        long start = System.currentTimeMillis();
        
        System.out.println("Reading file...");
        Scanner input = new Scanner(new File("smallmoby.txt"));
        while (input.hasNext()) {
            String word = input.next();
            // if (!words.contains(word)) {
                words.add(word);
            // }
        }
        
        long end = System.currentTimeMillis();
        long elapsed = end - start;
        
        System.out.println("The file has " + words.size() + " words.");

        // eliminate all plural words  (end with S)
        // need to use an iterator instead of a for-each loop
        Iterator<String> itr = words.iterator();
        while (itr.hasNext()) {
            String word = itr.next();
            if (word.endsWith("s")) {
                itr.remove();  // crash
            }
        }

        System.out.println("The file has " + words.size() + " words.");
        System.out.println("Took " + elapsed + " ms.");
    }
}