Java模式匹配器问题

Question

I'm looking for a pattern to check for boolean terms but only in certain conditions:

• If they exist by them selves
String a = "AND";
String b = "\\tand";
String c = "and ";
• If they are not part of a word or phrase
String d = "This or that";
• Terms or phrases to ignore:
String e = "band";
String f = "L'or"
String g = "can do";

The code I have so far only find them if they have spaces before and after the delimiter and any sort of adjustment breaks what progress I have. I used this page as reference but still no dice. I have tried using both find() and matches() but find seems to be too broad in its scope and matches doesn't seem broad enough. Any ideas?

final static Pattern booleanTerms = Pattern.compile("(.*)(( OR )|( or )|( NOT )|( not )( AND )|( and ))(.*)");

public static void main(String[] args) {

Set<String> terms = new HashSet<String>();
terms.add(" OR"); //false
terms.add("or "); //false
terms.add("OR"); // false
terms.add(" or "); //true
for (String s : terms) {
    System.out.println(findDilims(s));
} // end for loop

} // end main method

public static boolean findDilims(String s) {
    Matcher matcher = booleanTerms.matcher(s);
    if (matcher.matches()) {
      return true;
    } else {
      return false;
    }
} // end method

Answer 1

The reason that you are getting false for " OR", "or ", and "OR" is because your pattern is explicitly looking for boolean terms with a space both before or after: eg ( OR ) looks for " OR " .

Instead of requiring spaces before and after to ensure that each boolean term is a word, you probably want to use word boundaries instead:

Pattern.compile("\b(( OR )|( or )|( NOT )|( not )|( AND )|( and ))\b");

You can use \\s* to add optional whitespace to the beginning and end of your regex. That way " OR \\t " will also match.

Pattern.compile("\s*\b(( OR )|( or )|( NOT )|( not )|( AND )|( and ))\b\s*");

matcher.matches() should now work fine.

Answer 2

You said you want to find them only if they are by themselves and not as part of a phrase. Then you don't want to start and end your pattern with (.*) .

It seems you also want to find them even if there is whitespace around them. Then you would need to start and end your patter with \\s* . You also want to find them even if there is no space before or after. Then you don't want to have space on your patterns, like for example in ( or ) .

And it seems you want this to be case insensitive so you might want to set that with (?i)

final static Pattern booleanTerms = Pattern.compile("(?i)(\s*)((or)|(not)|(and))(\s*)");

Answer 3

You need a character class either end of the terms as an alternation:

(?i)(^\s*|[^a-z]\s)(or|not|and)(\s[^a-z]|\s*$)

And you only need one line:

public static boolean findDilims(String s) {
    return s.matches(".*(?i)(^\\s*|[^a-z]\\s)(or|not|and)(\\s[^a-z]|\\s*$).*");
}