Java：Pattern and Matcher (轉)

Java: Pattern and Matcher

In addition to the regular expression methods that are available in the String class (see String Regular Expressions), there are two classes that are specifically user for regular expression matching.

• java.util.regex.Pattern

   precompiles regular expressions so they can be executed more efficiently. It also has a few utility functions. This same pattern can be reused by many 

  Matcher

  objects.

  Pattern pat = Pattern.compile(regexString);
        

• java.util.regex.Matcher

   objects are created from a Pattern object and a subject string to scan.     This class provides a full set of methods to do the sacnning.

  Matcher m = pat.matcher(subject);
        

Common methods

The following variables represent elements of their declared type in the prototypes below.

import java.util.regex.*;
. . .
    boolean b;       // may be used in if statement.
    int i;           // index or character position.
    int g;           // group number
    int n;           // number of groups
    CharSequence cs; // effectively either String or StringBuffer
    String s;        // Subject string.
    String regex;    // Regular expression string.
    StringBuffer sb; // Used to build up string from repeated scans.
    Pattern p = Pattern.compile(regex); // Compiles regular expression into Pattern.
    Matcher m = p.matcher(s); // Creates Matcher with subject s and Pattern p.
      

Result	Method	Description
Creating a Pattern
p =	Pattern.compile(regex);	Creates Pattern based on regex. May throw  PatternSyntaxException .
p =	Pattern.compile(regex, f);	As above. Flag f can be  Pattern.CASE_INSENSITIVE , ....
Finding pattern matches
b =	m.find();	True if pattern can be found in the subject string. First call starts trying at beginning of string. Subsequent calls start after last character previously matched, which makes it good for a while loop.
b =	m.find(i);	True if pattern can match somewhere at or after position i.
b =	m.matches();	True if pattern matches entire subject string.
b =	Pattern.matches(regex, s);	As above, but less efficient if regex used more than once.
b =	m.lookingAt();	True if pattern matches starting at first char.
Getting results of last pattern match. Corresponds to group 0, then entire match.
s =	m.group();	String which was matched.
i =	m.start();	Index of first character of match.
i =	m.end();	Index of last character plus 1 of match.
Getting group results of last match
s =	m.group(g);	String which was matched by group  g .
i =	m.start(g);	Index of first character of group g.
i =	m.end(g);	Index of last character plus 1 of group g.
n =	m.groupCount();	Number of groups that were matched.
Misc
m =	m.reset();	Resets Matcher  m  so that next find starts at beginning.
m =	m.reset(cs);	Resets Matcher  m  to match subject  cs .
Replacing text
s =	m.replaceFirst(rep);	Returns string which is subject with first match replaced by  rep .
s =	m.replaceAll(rep);	Returns string with all matches replaced by  rep .
Building replacement in a StringBuffer
m =	m.appendReplacement(sb, s);	When it matches, (1) everything skipped before the match is appended to the StringBuffer sb, then (2) it appends s, with "$n" replaced by group n.
sb =	m.appendTail(sb);	Appends last part that m didn't match to StringBuffer. Useful after loop calling appendReplacement()  to finish.

Other Resources

• Strings: Compile Regular Expressions for Efficiency at www.javalobby.org
• Finding Comments in Source Code Using Regular Expressions
• The Regex Coach - interactive regular expressions