Regular Expressions

Regular expressions are a common way to describe text strings for searching. Many variants exist and we describe here the variant, which is used by Scanner.

A regular expression is a string in which some characters (metacharacters) have special meanings. The following characters are metacharacters:

* + ? [] ^ .

We can see a regular expression as a compact notation for a variety of ordinary strings (no special characters). We say that the expression match all strings in this set. We are content with a few examples:

  
• "a[dp]a" matches the two strings "ada" and "apa" (one must choose one of the options in brackets).   
• "a[d-g]a" matches four strings: "ada", "aea", "afa" and "aga" (one may describe options in a range).   
• "a[0-9]" matches the ten strings "a0", "a1", ... "a9" (range may comprise a range of letters or a range of numbers).   
• "ab+a" matches an infinite number of strings: "aba", "abba", "abbba", ... (+ means one or more occurrences of the previous item).   
• "a[bc]+d" also matches infinitely many strings: all consisting of an a followed by a non-empty sequence of b and c letters and the finally the letters d (the previous element is here is "b or c").   
• "a[bc]*d" matches the same strands as in the previous examples in addition to "ad" (* allows zero occurrences, + requires at least one).   
• "[0-9]+" matches all non-empty sequences of digits (i.e. all integers that are unsigned).   
• "a[^bc]*d" matches any string that begins with a, ends with d and where none of the intervening signs are b or c ([^] means   "none of the contained characters").   
• "a.*d" matches any string that starts with a and ends with d (. means "any character").   
• "abc" matches only the string "abc" (no metacharacter means a regular expression that matches itself).   

The method findInLine in Scanner takes a regular expression as a parameter and is looking for a matching substring in the current row. If it finds one, it returns it and the Scanner's current position is moved immediately after the matched string. If no matching string is found, it returns null and the current position is unchanged (i.e. you have searched forward on the line without success; leftover of the line being looked at is regarded as unread).   

A regular expression can match multiple substrings in a line: if the current line is "<tt>pattern</tt>", then the expression "<.*>" will match both the tag <tt> as well as the entire line. The rule is then that findInLine returns the longest matching string.