Last updated 2017 March 2

Regular expresions

Try the regular expression practice program.

Regular expression symbols as documented by developer.mozilla.org the developer of Firefox. Gives an explanation and examples for all the regular expression symbols. Here is the subset of symbols to know and be familiar with.

Here are a couple of interesting web sites discussing common regular expressions. Regular-expression.info and 8 Regular Expressions You Should Know - Code Tuts

Basic definition

To start you have an alphabet of characters.
For example the characters that you can type on a keyboard.

Putting a sequence of characters together you have a string.
For example "This is a string".

You can then have a set of strings; for example {"ax", "bx" "cx", "dx"}.

A regular expression describes a set of strings over an alphabet of characters.
For example {"ax", "bx" "cx", "dx"} can be described by the regular expression [abcd]x
While {"xa", "xb" "xc", "xd"} can be described by the regular expression x[abcd]

Regular expression symbols

Regular expressions consist of a string of symbols, where the following symbols are used in regular expressions.

Characters represent themselves, except for special characters, some of which are: [ ] { } ( ) . + * $ ? \ | ^
Special characters are written as "\c", where c is a special character. The character \ is an escape character, that means take the following character literally instead of its special meaning. For example the string of special characters in the item above would be represented as follows.
\[ \] \{ \}  \. \+ \* \$ \? \\ \| \^
[aeiou] square brackets are used to designate a set of characters any one of which could match. Ranges [e-p], [2-5] are accepted.
. any character, except newline
[^ceh] c, e, and h are characters that cannot be in a string, Ranges are accepted.
^ the beginning of a string
$ the end of a string
| or, a choice of characters in a string — (sa | pe) — use parenthesis to denote subexpressions
c* zero or more occurences
c+ one or more occurences
c? zero or one occurence
c{p} Matches the character c exactly p times
c{p,q} Matches the character c times between p and q inclusive.
(x) maches the pattern x and remembers the match. The parenthesis are called capturing.
Try /(foo) (bar) \2 \1/ on "foo bar bar foo"
(?:x) matches x but does not remember the match.
- /(foo) (?:a|b) (bar) \2 \1/ maches "foo a bar bar foo" and "foo b bar bar foo"
  – now remove the "?:" and to match the pattern you will have to change "\2" to "\3"
  /(foo) (a|b) (bar) \3 \1/ matches "foo bar a bar foo" and "foo bar b bar foo"
Often used with a following {p} or {p,q} because the parenthesis are only used for grouping, as in the following examples.
- (?:foo){3} — try on "foo foofoo foofoofoo"
- (?:foo){1,2} — try on "foo foofoo foofoofoo" and "foofoo bar fooo bar"
x(?=y) matches x only if followed by y. This is call lookahead.
Try "Jack (?=Sprat)" and "Jack (?=(Sprat|Dempsey))" on "Jack Sprat and Jack Frost and Jack Dempsey".
x(?!y) matches x only if not followed by y.
Try "Jack (?!Sprat)" and "Jack (?!(Sprat|Dempsey))" on "Jack Sprat and Jack Frost and Jack Dempsey".
\b macthes a word boundary
\B matches a non-word boundary
\d matches a digit
\D matches a non-digit
\n \r \t match a newline, a carridge return, and a tab
\s matches a single whitespace charater including newline, carridge return and tab.
\S matches a single non-whitespace character
\w matches any word character (alphanumeric including underscore)
\W matches any non-word character.
\xhh matches the character with the hexadecimal code hh
\uhhhh matches the character with the hexadecimal code hhhh (Unicode characters have 16 bits)

Regular expressions may also include the following search modifiers.

i — case insensitive search
g — global search; return all matches, not just the first one
m — treat a multi-line string as multiple lines and not as one line.

Look at examples in developer.mozilla.org webpage on regular expressions.

Try /Chapter \d+\.\d*/ — Try on "Read Chapter 4.32, Chapter 15. and Chapter 3 and 4."
The following pattern is one line, the browser puts in a "view break"
/(a?b?c?d?e?f?g?h?i?j?k?l?m?n?o?p?q?r?s?t?u?v?w?x?y?z?(?![a-z])|^a?b?c?d?e?f?g?h?i?j?k?l?m?n?o?p?q?r?s?t?u?v?w?x?y?z?(?![a-z]))/
matches any word with its letters in alphabetic order — almost, beknow, biopsy, chintz, ghosty.
/[^\s,aeiou]*a[^aeiou]*e[^aeiou]*i[^aeiou]*o[^aeiou]*u[^aeiou,]*/
matches any word with the five vowels in alphabetic order. — abstemious, abstemiously, facetious, fracedinous, majestious

Functions that use regular expressions

The following JavaScript functions use regular expressions for a pattern; and they can also use a string.

string.match(pattern) — returns an array (list) of all the matches of the pattern in the string.
string.search(pattern) — returns the index of the match in the string or -1 if the search fails.
string.replace(pattern, replacementPattern) — changes the string by replacing all instances of the pattern match with the replacementPattern
string.split(pattern) — returns an array of items that were separated by the pattern.

Creating regular expressions

Try the examples in a script followed by console.log(Result) statements.

In JavaScript a literal regular expression is denoted using the "/" character.

var theRegularExpression = /db+d/g;

        var theString = "cdbbdbsbz";

        var Result = theString.replace(theRegularExpression, "123");

An alternative is to use the regular expression constructor.

var theRegularExpression = new RegExp("db+d","g");

        var theString = "cdbbdbsbz";

        var Result = theString.replace(theRegularExpression, "123");

Use regular expressions to swap names.

var theRegularExpression = /(\w+) (\w+)/;

        var theString = "Leila Anderson";

        var Result = theString.replace(theRegularExpression, "$2, $1");

Previous example with '\' as the separator. Need to escape the '\' in the regular expression and in the string.

var theRegularExpression = /(\w+)\\(\w+)/;

        var theString = "Leila\\Anderson";

        var Result = theString.replace(theRegularExpression, "$2, $1");

An alternative is to use the regular expression constructor. In JavaScript you have to use '\\' in the string to get a single '\' in the pattern.

var theRegularExpression = new RegExp("(\\w+) (\\w+)","");

        var theString = "Leila Anderson";

        var Result = theString.replace(theRegularExpression, "$2, $1");

Previous example with '\' as the separator. Now need four '\' in the regular expression string to leave two '\' for the regular expression interpreter.

var theRegularExpression = new RegExp("(\\w+)\\\\(\\w+)","");

        var theString = "Leila\\Anderson";

        var Result = theString.replace(theRegularExpression, "$2, $1");

Putting escape characters into a string

The function escapeRegExp(string) – puts the escape character in front of every character that needs to be escaped in string to prevent it from being a special character in a regular expression

function escapeRegExp(string){

          return string.replace(/([.*+?^${}()|[\]\\])/g, "\\$&");

        }

        
  // Use the function

        var theString = "[ ] { } ( ) . + * $ ? \\ |";    
        // String in which to escape characters


        if (escapeRegExp("[ ] { } ( ) . + * $ ? \\ |") ===   // Compare result

          "\\[ \\] \\{ \\} \\( \\) \\. \\+ \\* \\$ \\? \\\\ \\|") {

          console.log("Escaped");

        }

        else { console.log("Still a prisoner"); }


        console.log(theString.length + "  " + escapeRegExp(theString).length);

        console.log(">>>" + theString + "<<<");

          console.log(">>>" + escapeRegExp(theString) + "<<<");

More complex examples

Try the examples in a script followed by console.log() statements for the variables.

Given a string of name-pairs, firstName lastName, with each pair separated by a semi-colon, and with arbitrary white space create a string of names with the first and last names swapped.

var names = "Harry Trump ; Fred Barney; Helen Rigby ; Bill Abel ; Chris Hand ";
 
        var namePattern = /(\w+)\s+(\w+)/g; // firstName whiteSpace lastName

        var newNames = names.replace(namePattern, "$2, $1");

Given a string of name-pairs, firstName lastName, with each pair separated by a semi-colon, and with arbitrary white space create a list of names with the first and last names swapped.

var names = "Harry Trump ; Fred Barney; Helen Rigby ; Bill Abel ; Chris Hand ";
 
        var splitPattern = /\s*;\s*/; // All white space before and after a ';'

        var nameList = names.split(splitPattern);

        var namePattern = /(\w+)\s+(\w+)/;  // firstName whiteSpace lastName

        var namesSwappedList = [];   // Create an empty list

        var index = 0;

        while (index < nameList.length) {

          namesSwappedList[index] = nameList[index].replace(namePattern, "$2, $1");

          index += 1

        }

An alternate way of doing the previous problem.

var names = "Harry Trump ; Fred Barney; Helen Rigby ; Bill Abel ; Chris Hand ";
 
        var namePattern = /(\w+)\s+(\w+)/g;  // firstName whiteSpace lastName

        var newNames = names.replace(namePattern, "$2, $1");
        
        var splitPattern = /\s*;\s*/; // All white space before and after a ';'

        var namesSwappedList = newNames.split(splitPattern);

A phone number consists of a three-digit area code followed by a three-digit exchange followed by a four-digit local number.

Variations are to permit the area code to be with parenthesis, e.g. (905), and to permit any of the characters ' '(space), '.' or '-' to separate the three parts, e.g. 123 456 9876, but the same character must be used in both places.

var phoneNumberPattern = /(?:\d{3}|$\d{3}$)([\-\.\s])\d{3}\1\d{4}/;
An example web page is here.

it is important to document complex regular expressions.

(?:\d{3}) matches 3 digits, does not remember the match
$\d{3}$ matches "( 3 digits )"
(?:\d{3}|$\d{3}$) matches either (1) or (2),does not remember the match
([\-\.\s]) matches a dash, period or space character and remembers it
\d{3} matches the next 3 digits
\1 matches the dash, period or space matched previously in (4)
d{4} matches the last 4 digits

Look at the JavaScript in the regular expression practice program.