Pattern Matching

Intro

• Classes in the java.util.regex package provide regular expressions support.
• The Pattern class is used to store a regex expression - the regex has to be "compiled."
• The Matcher class is used to start the regex engine to perform match operations.
• Basic example

import java.util.regex.*;

public class RegexTest1 {

	public static void main(String[] args) {
		
		Pattern p = Pattern.compile("lazy"); //The pattern to search for
		Matcher m = p.matcher("The quick brown fox jumps over the lazy dog"); //The source against which to match the pattern
		boolean found = false;
		while(m.find()) {
			System.out.println("Match found at " + m.start() + "," + m.end()); //Will print : Match found at 35,39
			found = true;
		}
		
		if(!found) {
			System.out.println("No match found");
		}
	}

}

• Thumb rule: Regex matching runs from left to right and once a source character has been consumed, it cannot be reused.
• In the below example, it will match the pattern "aba" starting at 0 and 4, but not at 2 since they are consumed during the match starting from 0.

import java.util.regex.*;

public class RegexTest2 {

	public static void main(String[] args) {
		
		Pattern p = Pattern.compile("aba");
		Matcher m = p.matcher("abababa");
		boolean found = false;
		while(m.find()) {
			System.out.println("Match found; starting at pos : " + m.start());
			found = true;
		}
		
		if(!found) {
			System.out.println("No match found");
		}
	}

}

Metacharacters

• Regex keywords that have special search meaning.
• \d - Matches a digit
• \D - Matches a non-digit equivalent to [^\d]
• \s - Matches a whitespace char
• \S - Matches a non-whitespace char
• \w - Matches a word char (letters/digits or _)
• \W - Matches a non-word char
• Dot - "." metacharacter matches any character

public class RegexTest3 {

	public static void main(String[] args) {
		
		Pattern p = Pattern.compile("\\d");
		Matcher m = p.matcher("The 15th of August");
		boolean found = false;
		while(m.find()) {
			System.out.println("Match found; starting at pos : " + m.start());
			found = true;
		}

                // Match found; starting at pos : 4
                // Match found; starting at pos : 5
		
		if(!found) {
			System.out.println("No match found");
		}
	}

}

• Character classes
• The [] notation is used to define a pattern that represents a set of characters. e.g:
• The search will match any of the chars defined within [] that is the "OR" operator will be used.
  • [abc] - Only a's or b's or c's
  • [a-f] - Search for a,b,c,d,e,f chars
  • [a-fA-F] - small and caps
  • [^aeiouAEIOUS] - no vowels

Quantifiers

• Used to specify the number of occurrences of a search pattern
• * - Zero or more occurrences
•  ? - Zero or one occurrence.
• + - One or more occurrence.

• The above three are greedy quantifiers.

• Example the pattern abc(\d)* will match -
  • abc0
  • abc13423
  • abc - since * means 0 or more
  • abcdef - for the similar reason as above

• It won't match -
  • ab211 (doesnt start with abc)
  • abcs (doesnt have a digit after abc)

Greedy Quantifiers

• Greedy quantifiers will try to look at the entire source data while trying to determine a match.

See example below:

public class Greedy {

	public static void main(String[] args) {
		String greedyPattern = ".*xx";
		String reluctantPattern = ".*?xx";
		String source = "yyxxxxyxx";
		
		Pattern gp = Pattern.compile(greedyPattern);
		Matcher gm = gp.matcher(source);

		while (gm.find()) {
			System.out.println("Greedy Match found ! Starts at : " + gm.start()
					+ ", Matched portion : " + gm.group());
			
		}
                //Will print:
                //Greedy Match found ! Starts at : 0, Matched portion : yyxxxxyxx

		Pattern rp = Pattern.compile(reluctantPattern);
		Matcher rm = rp.matcher(source);
		
		while (rm.find()) {
			System.out.println("Reluctant Match found ! Starts at : " + rm.start()
					+ ", Matched portion : " + rm.group());
			
		}
                //Will print:
                //Reluctant Match found ! Starts at : 0, Matched portion : yyxx
                //Reluctant Match found ! Starts at : 4, Matched portion : xx
                //Reluctant Match found ! Starts at : 6, Matched portion : yxx
	}

}

Tokenizing

• Tokenizing for small pieces of data can be done by the String.split() method.
• For advanced Tokenizing, using the Scanner class is the best choice.
• The scanner class can accept various forms of input such as files, streams or Strings.
• Tokenizing is done within a loop, so that the process can be exited once any conditions are met.
• Tokens can be converted to their primitive types automatically.
• In example below a scanner tokenizes a string containing integers. The default delimiter of a scanner is a whitespace character.

public class ScannerTest1 {
	
	private static String source = "M 78 P 85 C 92 E 66 B 88";

	public static void main(String[] args) {
		
		List<Integer> scores = new ArrayList<Integer>(); 
		
		Scanner scanner = new Scanner(source);
		
		while(scanner.hasNext()) {
			if(scanner.hasNextInt()) {
				int score = scanner.nextInt();
				scores.add(score);
			} else {
				scanner.next(); 
			}
		}
		
		Collections.sort(scores);
		
		System.out.println(scores);
	}
}

• Another example, where a regex is being used as a delimiter to the scanner:

import java.util.*;

public class ScannerTest2 {
	
	private static String source = "ABC = 322, DEF = 343, GHI = 522, KLM = 747"; 

	public static void main(String[] args) {
		
		Scanner scanner = new Scanner(source);
		scanner.useDelimiter(",\\s*");
		
		Map<String, Integer> nameValueMap = new HashMap<String, Integer>();
		
		while(scanner.hasNext()) {
			String token = scanner.next();
			Scanner lineScanner = new Scanner(token);
			lineScanner.useDelimiter("\\s=\\s");
			String name = null;
			int value = 0;
			while(lineScanner.hasNext()) {
				if(lineScanner.hasNextInt()) {
					value = lineScanner.nextInt();
				}  else {
					name = lineScanner.next();
				}
			}
			nameValueMap.put(name, value);
		}
		
		System.out.println(nameValueMap);
		
	}
}