Matching patterns in strings is important in just about all Web applications that deal with data. PHP has a large number of pattern matching functions and extensions, tightly integrated into the language itself. The easiest pattern in a string to match is a single character. This can be accomplished in PHP as follows:

01 <?
02 $c = 'a';
03 $str = 'string with a';
04 $found = 0;
05 for($i=0;$i<strlen($str);$i++) {
06       if($str[$i] == $c) {
07 		$found = 1;
08 		break;
09 	 }
10 }
11 ?>

In this basic script, we loop through each character of $str, testing it against the character we are looking for, $c. If we find it, we set $found to true and break the loop.

Conveniently, PHP has a function which does exactly this. The following does the same:

01 <?
02 $found = (strpos($str,$c) === false ? 0 : 1);
03 ?>

The function strpos() searches for $c in $str. It returns the offset of the first occurrence of $c, or false if $c isn't found. Notice that we use '===' to ensure that false is being returned and not 0, which is the first character of $str.

Another common requirement is to find the portion of a string matching a number of characters. For example, say you wanted to read a string containing numeric data up until a non-numeric character appeared. Consider the following:

01 <?
02 $nums = "1234567890";
03 $str = "340872 * 10";
04 $numlen = strspn($str,$nums);
05 $numeric = substr($str,0,$numlen - 1);
06 ?>

Advanced pattern matching

Widely used on UNIX systems and in many programming languages, regular expressions are probably the most sophisticated all-purpose textual pattern matching mechanisms. A regular expression is a script which denotes a pattern.

The most elementary regular expressions are 'literals', i.e., characters which are interpreted literally. Consider the following:

01 <?
02 $regs = array();
03 if(ereg("abc","a string with abc",$regs)) {
04     echo "Pattern matched: {$regs[0]}\n";
05 } else {
06     echo "No pattern matched\n";
07 }
08 ?>

The ereg() function matches patterns in the first argument within the string passed as the second argument. Matches for this type of pattern are stored in the array $regs, passed as the third argument.

This kind of pattern matching can easily be accomplished with the functions already covered; what those functions cannot do is match wild cards. Consider the following:

01 <?
02 $regs = array();
03 if(ereg("wildcard: (.)","The wildcard: x",$regs)) {
04     echo "The wildcard matched: {$regs[1]}\n";
05 } else {
06     echo "No pattern matched\n";
07 }
08 ?>

The pattern used this time is decidedly different. There are two distinct parts. The first, "wildcard: " (note the trailing space) are treated literally. The second part, "(.)", is used to match the character "x" in the string. The stop (".") matches any character except a newline. The parenthesis instructs ereg() to place the matched character (that is, the character matching ".") to be placed in $regs. The first matched parenthesis starts at $regs[1].

01 <?
02 $regs = array();
03 if(ereg("wildcard: (.)","The wildcard: x",$regs)) {
04     echo "The wildcard matched: {$regs[1]}\n";
05 } else {
06     echo "No pattern matched\n";
07 }
08 ?>

Pattern matching can be vastly enhanced. When an asterisk ("*") is appended to a pattern, the regular expression matches zero or more instances of that pattern. For example, a regular expression "abc*" will match "ab" (remember: zero or more instances), "abc", "abcc", "abccccc" and so on. Notice that "abc*" does not match zero or more instances of "abc", but only "ab" followed by zero or more instances of the character "c".

To match "abc" zero or more times, the regular expression would be: "(abc)*". In this case, "abc" is enclosed in parentheses and is treated as an "atom" or "single unit".