PHP: Regular expressions

A regular expression, or rational expression, allows you to define a general pattern, called regular expression pattern, by means of particular characters, representative of a set of very varied Strings.

If you define, for example, as a characteristic of a word model that the

first and third letter of a word must be a "p", quantity of words of a

English dictionary meet this criteria, such as pipe, etc. These

patterns allow to search in a text if there are occurrences of a word or

find all the words that meet such a particular criterion, as you have just

see.

They can also make it possible to check if an input made by a visitor of the site is

conform to the commonly expected format, such as that of an e-mail address, and in the

otherwise, display an error message. Without this type of verification, any contact

would be impossible between a customer who identified himself with an address containing an error format and a commercial site that saves email addresses to send

an order confirmation, for example.

You will see successively how to write representative patterns and then the

various functions offered by PHP to perform checks on strings of

characters. In practice, these strings come mostly from inputs

made in a form.

Definition of a basic pattern

Writing regular expression patterns is the most forbidding part of the job

coding. Great care is required in their writings as this conditions the quality

of the result. Patterns are always contained in character strings, and so

in quotation marks, and must be framed by a / (slash) character at the beginning and end

chain.

Search for one or more characters

To find the presence of a particular character, it is sufficient to include it between

single or double quotes. To search for the @ character, you write the model:

$ model = "/ @ /";

To check if a string contains at least one character from a list, you

list all these characters in square brackets.

To search for one or more xyz characters, you have the following reason:

$ model ="/[xyz]/";

With the same syntax, you can define as a pattern an interval of letters or

figures.

For example, if you write:

$ model = "/ [a-z] /";

you are looking for a word of at least one character containing any letters

tiny.

Similarly, the following reason:

$ model = "/ [A-Z] /";

look for any group of capital letters.

The following :

$ model = "/ [0-9] /";

search for the presence of at least one digit between 0 and 9.

There are predefined character classes, which avoid having to create yourself

the sets of characters sought. These classes are summarized in the table below:

The classes of the table are usable like the other reasons by writing them in a

string delimited by quotation marks.

For example, you write the following template:

$ model = "/ [[: alpha:]] /";

To search for lowercase or uppercase alphabetic characters.

The following model:

$ model = "/ [[: digit:]] /";

allows you to search for the presence of numbers.

Finding a precise String

To find the presence of a specific word in a string, create the pattern by writing

the word in quotation marks. The search is validated if the word is exactly in the

chain analyzed and respecting the case.

The following reason:

$ model = "/ Paris /";

check if the word Paris is present in the parsed string. The chain Visit in Paris is

therefore conform to the model but not the chain Visit to Paris.

To detect the presence of at least one of two channels, you must use

the OR logical operator, symbolized by the pipe (|) character placed between the words to search.

The following reason:

$ model = "/ \ com | \ .net /.";

search for .com or .net strings. The address www.php.net is validated but not

www.w3c.org.