Quantifiers and character classes

Step 3's regexes described texts of a fixed length. Most formats have parts that repeat: a postcode is five digits, a PIN four to six, and a name any number of letters. Two more tools describe them: quantifiers and character classes.

Quantifiers

A quantifier goes after a character, or after a group, and says how many times it may come:

Without parentheses, a quantifier repeats only the one character right before it. ha+ is an h followed by one or more a's, and (ha)+ is ha, once or more:

System.out.println("haaa".matches("ha+"));
System.out.println("haaa".matches("(ha)+"));
System.out.println("hahaha".matches("(ha)+"));
System.out.println("gal".matches("go*al"));

true
false
true
true

go*al allows zero o's, so gal matches it, but not go+al.

Character classes

Square brackets make a character class. It matches one character, any of the ones inside: [aeiou] is one vowel. A dash between two characters gives a range: [0-9] is one digit, [a-z] one lowercase letter, [A-Z] one capital letter, and [a-zA-Z] a letter of either kind.

With a quantifier, a class describes a whole format. A US ZIP code is five digits, sometimes followed by a dash and four more digits (the ZIP+4 form, such as 10001-4356):

String zip = "[0-9]{5}(-[0-9]{4})?";
String[] codes = {"90210", "10001-4356", "9021", "90210-", "9021O"};
for (String code : codes) {
    System.out.println(code + " " + code.matches(zip));
}

90210 true
10001-4356 true
9021 false
90210- false
9021O false

• [0-9]{5} is five digits.
• (-[0-9]{4})? is a group of a dash and four digits, and the ? makes the whole group optional. A dash outside square brackets is just a dash.
• 90210- has the dash, but not the four digits that must come with it.
• 9021O ends with the capital letter O, which isn't in [0-9].

Times of day

Sometimes a class isn't enough by itself, and a group with | does the rest. A time on a 24-hour clock is written like 07:45 or 23:05. The hours go from 00 to 23: a 0 or a 1 followed by any digit, or a 2 followed by a digit from 0 to 3. The minutes go from 00 to 59:

String time = "([01][0-9]|2[0-3]):[0-5][0-9]";
System.out.println("07:45".matches(time));
System.out.println("23:05".matches(time));
System.out.println("24:00".matches(time));
System.out.println("7:45".matches(time));
System.out.println("12:60".matches(time));

true
true
false
false
false

The group holds the two ways to write an hour, and the colon stands for itself.

A few more

• A dot . matches any one character, so .* matches any text on one line: "b.*" matches every one-line string that starts with b.
• \\d in a Java String is a shorter way to write [0-9], so "\\d{5}" is five digits.
• To match a real dot, write "\\.". Module 16's split takes a regular expression too, so split(".") splits at every character, not just at the dots (a dot means any character), but split("\\.") splits only at the dots.