Regex to fetch the correct java class name

Question

For some reason, I want scan the content of java file(e.g. TagMatchingInterface.java) and fetch the class name(TagMatchingInterface) via regex, but my regex match the incorrect class name as there are some key words(class/interface/enum) hiding in the comment:

/**
 *
 * @author XXXX
 * Introduction: A common interface that judges all kinds of algorithm tags.
 * some other comment
 */
public class TagMatchingInterface 
{
  // content
  public class InnerClazz{
    // content
  }
}

here is my pattern:

public Pattern CLASS_PATTERN = Pattern.compile("(?:public\\s)?(?:.*\\s)?(class|interface|enum)\\s+([$_a-zA-Z][$_a-zA-Z0-9]*)");
....
Matcher matcher = CLASS_PATTERN.matcher(content);
if (matcher.find()) {
   System.out.println(match.group(2));
}

Any idea about my regex?

For some reason What reason? Java is a compiled language, at runtime you don't typically have the source. — Elliott Frisch
– Elliott Frisch, Commented May 24, 2016 at 3:05
I want use my custom classload to load class from MySQL, thus have to scan the content as a multi-line string — vash_ace
– vash_ace, Commented May 24, 2016 at 3:14

Ro Yo Mi · Accepted Answer · 2016-05-24 04:25:34Z

7

Description

(?<=\n|\A)(?:public\s)?(class|interface|enum)\s([^\n\s]*)

Regular expression visualization

This regex does the following:

allow the string to start with public or not
be a class or interface or enum
capture the name

Note, I recommend using the global and case insensitive flags

Example

Live Example

https://regex101.com/r/vR0iK3/1

Sample Text

/**
 *
 * @author XXXX
 * Introduction: A common interface that judges all kinds of algorithm tags.
 * some other comment
 */
public class TagMatchingInterface 
{
  // content
  public class InnerClazz{
    // content
  }
}

Sample Matches

[0][0] = public class TagMatchingInterface
[0][1] = class
[0][2] = TagMatchingInterface

Capture groups:

group 0 gets the entire match
group 1 gets the class
group 2 gets the name

Explanation

NODE                     EXPLANATION
----------------------------------------------------------------------
  (?<=                     look behind to see if there is:
----------------------------------------------------------------------
    \n                       '\n' (newline)
----------------------------------------------------------------------
   |                        OR
----------------------------------------------------------------------
    \A                        Start of the string
----------------------------------------------------------------------
  )                        end of look-behind
----------------------------------------------------------------------
  (?:                      group, but do not capture (optional
                           (matching the most amount possible)):
----------------------------------------------------------------------
    public                   'public'
----------------------------------------------------------------------
    \s                       whitespace (\n, \r, \t, \f, and " ")
----------------------------------------------------------------------
  )?                       end of grouping
----------------------------------------------------------------------
  (                        group and capture to \1:
----------------------------------------------------------------------
    class                    'class'
----------------------------------------------------------------------
   |                        OR
----------------------------------------------------------------------
    interface                'interface'
----------------------------------------------------------------------
   |                        OR
----------------------------------------------------------------------
    enum                     'enum'
----------------------------------------------------------------------
  )                        end of \1
----------------------------------------------------------------------
  \s                       whitespace (\n, \r, \t, \f, and " ")
----------------------------------------------------------------------
  (                        group and capture to \2:
----------------------------------------------------------------------
    [^\n\s]*                 any character except: '\n' (newline),
                             whitespace (\n, \r, \t, \f, and " ") (0
                             or more times (matching the most amount
                             possible))
----------------------------------------------------------------------
  )                        end of \2
----------------------------------------------------------------------

answered May 24, 2016 at 4:25

Ro Yo Mi

15k5 gold badges38 silver badges43 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

vash_ace Over a year ago

MvG Over a year ago

This doesn't match class if it's indented. It doesn't handle other class modifiers. And the look-behind is way more confusing that ^ combined with the MULTILINE flag would have been.

timlg07 Over a year ago

This would match the { in a case like public class TagMatchingInterface{ (no space or new line after the classname). So I would just go with (?<=\n|\A)(?:public\s)?(class|interface|enum)\s(\w+) (even shorter)

Ro Yo Mi Over a year ago

Shorter isn't always better. \w+ will match A-Z, a-z, 0-9, and _ characters. Java allows class names to include A-Z, a-z, 0-9, _, and $. Not sure why someone would include a dollar sign but we just turned on to Pedantic Drive. With that said, I'm sure the [^\n\s]+ would be overly greedy and could be scaled back to [^\n\s{]+ to prevent the capture of the open bracket, or just replace that with [A-Za-z0-9_$]+ since those are the only allowed characters in a class name.

MvG · Accepted Answer · 2016-05-24 05:10:17Z

1

… there are some key words(class/interface/enum) hiding in the comment:

Then get rid of all comments first. Suitable regular expressions should be fairly easy to write. I suggest you eliminate both kinds of comments (single line and multiple lines) simultaneously, in case there is text in one that looks like the start of the other.

Also get rid of all strings, while you are at it, since you might have a string in an annotation preceding the class.

"(?:public\\s)?(?:.*\\s)?(class|interface|enum)\\s+([$_a-zA-Z][$_a-zA-Z0-9]*)"

Checking for public serves little purpose, since the part without it would match just as well. In fact only the later parts will match if one of the class modifiers like final or abstract follows public.

So if you want to know whether the class is in fact public, you have to check for these, too. Which will be tricky since you may have annotations with parenthesized arguments nested to arbitrary depth. That's something a regular expression can't handle correctly.

What about classes containing non-ASCII letters in their name? What about unicode escapes in the input?

edited May 24, 2016 at 5:10

answered May 24, 2016 at 5:00

MvG

61.8k25 gold badges158 silver badges294 bronze badges

2 Comments

vash_ace Over a year ago

MvG Over a year ago

@vash_ace: Won't report public @Annotated class Foo as public.

Collectives™ on Stack Overflow

Regex to fetch the correct java class name

2 Answers 2

Description

Example

Explanation

4 Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Description

Example

Explanation

4 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related