Let's go a little further out; not too far - waist-deep will do. You're now ready to start exploring character class shortcuts. As we saw in the previous chapter, character class patterns match against a set of characters. We use these simple patterns a lot; in fact, you'll find that nearly every regex you use contains at least one character class pattern. It may not look like the character classes you have already seen, though; we use some character classes so often that most regex engines have built-in shortcuts. We'll explore those shortcuts in this chapter.
The most commonly needed character class is the class that represents any character; if you don't care whether a character is alphanumeric, punctuation, whitespace, some control character, or something else entirely, you need a character class that represents any character. That's where the . (a period) meta-character comes in. Try matching /./ against the following strings:
This text contains letters, numbers, punctuation,
whitespace, and even newline characters. 99988222.
Everything in it should match the /./ pattern.
Rubular should highlight every character in this text, including the spaces. The unhighlighted characters - you probably can't see this - are the newline characters at the end of each line.
By default, /./ does not match newline characters, which is useful: you typically don't want them. Use the /m (multiline) flag when . should match newlines; see the documentation for your language.
Even though . is a shortcut for a character class, it does not appear inside square brackets. A . inside square brackets is literal; if you want to match "any character", you must move the period outside the square brackets.
Two additional character class shortcuts that are frequently needed are \s for whitespace characters, and \S for non-whitespace characters. By definition, the whitespace characters are the space (' '), tab ('\t'), vertical tab ('\v'), carriage return ('\r'), line feed ('\n'), and form feed ('\f'). Thus, /\s/ is equivalent to:
/[ \t\v\r\n\f]/
while /\S/ is equivalent to:
/[^ \t\v\r\n\f]/
Consider these examples:
puts 'matched 1' if 'Four score'.match(/\s/)
puts 'matched 2' if "Four\tscore".match(/\s/)
puts 'matched 3' if "Four-score\n".match(/\s/)
puts 'matched 4' if "Four-score".match(/\s/)
if ('Four score'.match(/\s/)) {
console.log('matched 1');
}
if ("Four\tscore".match(/\s/)) {
console.log('matched 2');
}
if ("Four-score\n".match(/\s/)) {
console.log('matched 3');
}
if ("Four-score".match(/\s/)) {
console.log('matched 4');
}
The first three examples in each group all print a matched message because the given string contains a whitespace character; the last in each group outputs nothing since "Four-score" doesn't include whitespace.
Similarly:
puts 'matched 1' if 'a b'.match(/\S/)
puts 'matched 2' if " \t\n\r\f\v".match(/\S/)
if ('a b'.match(/\S/)) {
console.log('matched 1');
}
if (" \t\n\r\f\v".match(/\S/)) {
console.log('matched 2');
}
prints matched 1 since /\S/ matches each of the letters in 'a b', but does not print anything for the second match since all of the characters in the string are whitespace characters.
You can use \s and \S both in and out of square brackets. Outside square brackets, e.g., /\s/, \s stands for one of the whitespace characters. Inside square brackets, e.g., /[a-z\s]/, \s represents an alternative to the other members of the class. That is, the regex represents any character that is either a lowercase alphabetic character or a whitespace character.
The decimal digits, 0-9, and the hexadecimal digits, 0-9, A-F, and a-f, also show up often in character classes, so we have shortcuts for them:
| Shortcut | Meaning |
|---|---|
| \d | Any decimal digit (0-9) |
| \D | Any character but a decimal digit |
| \h | Any hexadecimal digit (0-9, A-F, a-f) (Ruby) |
| \H | Any character but a hexadecimal digit (Ruby) |
Try each of these shortcuts against the following strings:
Launch school
July 4th, 1776
0xABCDef12
As with \s and \S, you can use these shortcuts in or out of square brackets.
This last pair of shortcuts may be confusing. /\w/ matches "word characters", while /\W/ matches "non-word characters". At first glance, most people assume that word characters are alphabetic. In fact, the definition of word characters for \w is a bit broader than that; they include all alphabetic characters (a-z, A-Z), all decimal digits (0-9), and, oddly, an underscore (_). Avoid trouble and commit this to memory.
Try the /\w/ and /\W/ patterns against these strings:
Launch school
July 4th, 1776
one_word_two_words
Don't fence me in.
There is no simple shortcut for alphabetic character classes.
As with \d and \D and some other shortcuts on this page, you can use both \w and \W in or out of square brackets.
That's a wrap for character classes and the most basic building blocks of regex. Liberal use of these shortcuts help make your regex easy to type and improve readability. We have one more concept to cover before fully immersing ourselves: anchors.
Before you go, though, we have some exercises for you. In these exercises, use Rubular to write and test your regex. You don't need to write any code, though you may need to use IRB or the JavaScript console for some items.
Write a regex that matches any sequence of three characters delimited by whitespace characters (the regex should match both the delimiting whitespace and the sequence of 3 characters). Test it with these strings:
reds and blues
the lazy cat sleeps
There should be three matches in Rubular, but only two in Scriptular. This difference is caused by how the two apps deal with newline characters.
/\s...\s/
As expected, this regex matches and and cat, together with the spaces to either side of those words. What might be more surprising is that the also matches on Rubular; here, the newline between the first and second lines of text is a whitespace character.
Test the pattern /\s...\s/ from the previous exercise against this text (be sure to delete the previous text first):
Doc in a big red box.
Hup! 2 3 4
Observe that one of the three-letter words in this text match the pattern; it also matches 2 3. Why is it that this pattern doesn't include the three-letter words Doc, red, box, or Hup, but it does match 2 3?
Note that in all of these cases, the "match" is five characters long:
Doc doesn't match since Doc doesn't follow any whitespace.
big matches since it is three characters with both leading and trailing whitespace.
red doesn't match since the regex engine consumes the space character that precedes red when it matches big (note the trailing space). Once consumed as part of a match, the character is no longer available for subsequent matches.
box doesn't match since a period follows it.
Hup doesn't match since an exclamation point follows it.
2 3 matches since 2 3 is three characters long and it has both leading and trailing whitespace.
Write a regex that matches any four digit hexadecimal number that is both preceded and followed by whitespace. Note that 0x1234 is not a hexadecimal number in this exercise: there is no space before the number 1234.
Hello 4567 bye CDEF - cdef
0x1234 0x5678 0xABCD
1F8A done
There should be four matches (2 on Scriptular)
/\s\h\h\h\h\s/
/\s[\dA-F][\dA-F][\dA-F][\dA-F]\s/ig
The real surprise here may be that cdef and 1F8A are matches. If you followed the previous exercise, though, it shouldn't come as a surprise; cdef has a trailing white space character in the form of a newline, and 1F8A has a preceding white space that is a newline.
Note that the JavaScript solution cannot use \h, but needs to use [\dA-F] instead, or, equivalently, [0-9A-F].
The matches are 4567, CDEF, cdef, and 1F8A. On Scriptular, those last two numbers fail to match.
Write a regex that matches any sequence of three letters. Test it with these strings:
The red d0g chases the b1ack cat.
a_b c_d
There should be seven matches.
/[a-z][a-z][a-z]/i
This question was tricky in that it doesn't use any character class shortcuts; recall that there isn't one for letters. Note that /\w/ matches digits and underscores as well.
If you entered something different, check your work: Rubular should highlight The, red, cha, ses, the, ack, and cat if your regex is correct. Note in particular that neither d0g (dee-zero-gee) nor b1a (bee-one-ay) light up, nor do either of the underscored values.