Java string matching with wildcards

Question

I have a pattern string with a wild card say X (E.g.: abc*).

Also I have a set of strings which I have to match against the given pattern.

E.g.:

abf - false

abc_fgh - true

abcgafa - true

fgabcafa - false

I tried using regex for the same, it didn't work.

Here is my code

String pattern = "abc*";
String str = "abcdef";

Pattern regex = Pattern.compile(pattern);

return regex.matcher(str).matches();

This returns false

Is there any other way to make this work?

Thanks

Solution 1

Just use bash style pattern to Java style pattern converter:

public static void main(String[] args) {
        String patternString = createRegexFromGlob("abc*");
        List<String> list = Arrays.asList("abf", "abc_fgh", "abcgafa", "fgabcafa");
        list.forEach(it -> System.out.println(it.matches(patternString)));
}

private static String createRegexFromGlob(String glob) {
    StringBuilder out = new StringBuilder("^");
    for(int i = 0; i < glob.length(); ++i) {
        final char c = glob.charAt(i);
        switch(c) {
            case '*': out.append(".*"); break;
            case '?': out.append('.'); break;
            case '.': out.append("\\."); break;
            case '\\': out.append("\\\\"); break;
            default: out.append(c);
        }
    }
    out.append('$');
    return out.toString();
}

Is there an equivalent of java.util.regex for “glob” type patterns?
Convert wildcard to a regex expression

Solution 2

you can use stringVariable.startsWith("abc")

Solution 3

abc* would be the RegEx that matches ab, abc, abcc, abccc and so on.
What you want is abc.* - if abc is supposed to be the beginning of the matched string and it's optional if anything follows it.
Otherwise you could prepend .* to also match strings with abc in the middle: .*abc.*

Generally i recommend playing around with a site like this to learn RegEx. You are asking for a pretty basic pattern but it's hard to say what you need exactly. Good Luck!

EDIT:
It seems like you want the user to type a part of a file name (or so) and you want to offer something like a search functionality (you could have made that clear in your question IMO). In this case you could bake your own RegEx from the users' input:

private Pattern getSearchRegEx(String userInput){
    return Pattern.compile(".*" + userInput + ".*");
}

Of course that's just a very simple example. You could modify this and then use the RegEx to match file names.

Comments

[anubhava]

Can you show your code. What are all the wild cards you want to support?

[Ivar]

Why didn't the regex work? What did you try?

[Scary Wombat]

show us what you tried - maybe a simple fix

[Avinash Raj]

you mean abc.+ ?

[Scary Wombat]

why not just use String.contains ?

[Thomas]

That should be easy to do with regex, just replace the wildcards with the appropriate expression (e.g. .* for * and . for ?) and use String#matches().

[Thomas]

@ScaryWombat fgabcafa should not match but it would with "fgabcafa".contains("abc")

[Sunflame]

@Thomas that is not working but he can use the String::startsWith method. But the best is to use regexps.

[Scary Wombat]

@Thomas Oh sorry, I was reading the OP's question as that were his results, not what he was wanting - in that case startsWith

[Thomas]

@Sunflame why shouldn't that be working if he's using the correct regex? Assuming the wildcards can be anywhere (and probably multiple wildcards as well) you can't just use startsWith(), contains() or endsWith().

[Sunil Kumar B M]

Thanks.. but what I really want is linux style file matching like E.g: ls .sh will give all files ending with .sh or ls abc will give files starting with abc... I don't have control over the pattern or the actual strings...

[bkis]

Why don't you have control over the pattern? Does a user input the pattern?

[Sunil Kumar B M]

yes.. user give the pattern input.. and strings are a list of file names which I have to match

[bkis]

[a][b][c].* doesn't make sense. You are creating character classes with only one character in them. It's the same as abc.*

[Sunflame]

Yes you are right, I was checking with the IntelliJ built in Regexp check, and Idk why it didn't work with abc.*.

[Sunil Kumar B M]

that won't work... the input pattern provided is abc*, and basically, it could be anything like .sh, abc, I'm looking for linux style pattern matching

[Thomas]

Note that String#matches() doesn't require the expression to be wrapped in ^...$ since that's done implicitly anyways.

[Sunil Kumar B M]

@Thomas - noted.. pattern matching works without ^...$ wrapper also