Regex Woes

[Jim Longmore]

Hey all,

I'm trying to parse polynomials into their individual terms, e.g. "3x+2" into "3x" and "2", using regular expressions. However, I've no experience with Java regex. The code I've come up with below, after consulting the SDK is

                Pattern pattern = Pattern.compile("[*+|*-]");
String input = jtx.getText();
jtx.setText("");
Matcher match = pattern.matcher(input); 
jta.setText("");
System.out.println(match);
boolean found = match.lookingAt();
jta.append(String.valueOf(found)+"\n");
while(found)
{
jta.append(match.start()+"|" + match.end()+"\n");jta.append(input.substring(match.start(),match.end())+"\n");
match.replaceFirst("");
found = match.lookingAt();
}  
// jta is a JTextArea and jtx is a JTextField

The Match object is returning no match, although I've tested that expression using a 3rd party regex checker and it works. Am I using the Match or Pattern classes incorrectly? Any tips for regex in general are also welcome.

Thanks in advance,
Jim

[Scheepers de Bruin]

Well there is an easier solution:

String myPolynomal = "2x+3";

String terms[] = myPolynomal.split("[\+|-|\*|\/]");

+, * are meta characters, and need to be escaped (have a \ in front of them)

"Split myPolynomal on occurances of any in the set [ + OR - OR * OR / ]"

[Jim Longmore]

According to the documentation on the Java website, the String class doesn't have a split method, or have I missed something?

[Joanne Neal]

Originally posted by Jim Longmore:
According to the documentation on the Java website, the String class doesn't have a split method, or have I missed something?

You've missed the upgrade to Java 1.4

[Alan Moore]

Scheepers, I'm sorry to say your solution has several problems. Within a character class, most of the characters that usually have special meaning, don't. So there's no need to escape the '*', '+', '/' characters--but you should have escaped the '-' because character classes use it to define a range, as in "[a-z]" (but if it's the very first or last character in the set, the hyphen gets treated as a literal character). The '|' shouldn't be in there at all: the OR is implicit in a character class, so the pipe just gets treated a another literal character. Also, when you're creating a regex in a String literal, you have to use two backslashes to escape a character, not one.

Jim, the lookingAt() method means the regex matches the beginning of the target text, but not necessarily all of it. You probably want the find() method, which looks for the next match anywhere with the target. Here's a sample app to demonstrate:

import java.util.regex.*;
 
public class Test
{
  public static void main(String... args)
  {
    String target = "xxxyyyzzz";
 
    Pattern p1 = Pattern.compile("x+");
    Matcher m1 = p1.matcher(target);
    System.out.println(m1.find());          // true
    System.out.println(m1.lookingAt());     // true
    System.out.println(m1.matches());       // false
 
    Pattern p2 = Pattern.compile("y+");
    Matcher m2 = p2.matcher(target);
    System.out.println(m2.find());          // true
    System.out.println(m2.lookingAt());     // false
    System.out.println(m2.matches());       // false
 
    Pattern p3 = Pattern.compile("[xyz]+");
    Matcher m3 = p3.matcher(target);
    System.out.println(m3.find());          // true
    System.out.println(m3.lookingAt());     // true
    System.out.println(m3.matches());       // true
  }
} 

Also, input.substring(match.start(),match.end()) is exactly the same as match.group().

[Scheepers de Bruin]

I stand corrected

(So when is someone building a sample-code-compiler for the ranch to verify?)
=)

[Akshay Kiran]

well you will have problems with polynomials
how do you propose to match the constant coefficients with their respective powers of x?
for example
2x^2 + 3x + 5 might also come in as
3x + 2x^2 +5
do you allow only linear polynomials? and are you assured that the order will always be in the decreasing powers of x? and that they will be the most simplified form?
of course these don't come in as regex problems, but just thought I should mention them too.