Regex not working in PHP same as with C#

Question

For some reason the following regex which does work in .NET is not working in PHP:

// Line breaks not in real expression, just for clarity here to show sub-expressions:
$pattern = <<<REGEX
  "(\\"|[^"])*"
|
  '(\\'|[^'])*'
|
  [A-Za-z_][A-Za-z_\-\d]*
|
  [\-\+]?\d+(\.\d+)?
|
  [=<>!][=]
|
  [?:,()*\/\-\+!]
|
  \|\|?
|
\&\&?
REGEX;

Actual expression for testing:

$pattern = '/"(\\\"|[^"])*"|\'(\\\'|[^\'])*\'|[A-Za-z_][A-Za-z_\-\d]*|[\-\+]?\d+(\.\d+)?|[=<>!][=]|[?:,()*\/\-\+!]|\|\|?|\&\&?/';

$expr = <<<EXPR
something ? '<a href="example.com">', title, '</a>' : title
EXPR;

preg_match_all($pattern, $expr, $tokens);

Expected Output for $tokens[0]

0: something
1: ?
2: '<a href="something.com">'       // Please note that ' are part of token!
3: ,
4: title
5: ,
6: '</a>'
7: :
8: title

Actual Output for $tokens[0]

0: something
1: ?
2: a
3: href
4: something
5: com

For some reason most of the input is missing and the literal has been ignored. This works in .NET but for some reason is not working in PHP.

Just a thought: If you are trying to extract parts of your DOM, using something like php.net/manual/en/domdocument.loadhtml.php can make your life a lot easier when trying to extract information. — afuzzyllama
– afuzzyllama, Commented Aug 1, 2011 at 14:11
@afuzzyllama it isn't for analysing markup, the markup is just a string literal within my own simple expression language. It is for data binding and contains very basic syntax. My evaluator works fine when tokens are hard coded, but the regex isn't breaking the input expression down properly. — Lea Hayes
– Lea Hayes, Commented Aug 1, 2011 at 14:13
@afuzzyllama updated question with comment. the apostrophes ' are actually part of the required string token. — Lea Hayes
– Lea Hayes, Commented Aug 1, 2011 at 14:15
copy&paste -> Warning: preg_match_all(): Unknown modifier ']' please double-check the example script. — VolkerK
– VolkerK, Commented Aug 1, 2011 at 14:26
@VolkerK I am not getting that error, but I have put the exact version that I am using beneath the original sample. — Lea Hayes
– Lea Hayes, Commented Aug 1, 2011 at 14:30

ridgerunner · Accepted Answer · 2011-08-01 15:06:49Z

2

The problem is that you need to escape the regex backslash metacharacters. Also, it is best to write out complex regexes in free-spacing mode with lots of comments. Here is the PHP code you need:

<?php
$pattern = '%
      "(\\\\"|[^"])*"          # Either... a double quoted string,
    |                          # or...
      \'(\\\\\'|[^\'])*\'      # a single quoted string,
    |                          # or...
      [A-Za-z_][A-Za-z_\-\d]*  # an identifier.
    |                          # or...
      [\-\+]?\d+(\.\d+)?       # a number
    |                          # or...
      [=<>!][=]                # a comparison operator,
    |                          # or...
      [?:,()*\/\-\+!]          # a single char 
    |                          # or...
      \|\|?                    # a logical or numerical OR
    |                          # or...
    \&\&?                      # a logical or numerical AND
    %ix';

$expr = <<<EXPR
something ? '<a href="example.com">', title, '</a>' : title
EXPR;

$a = preg_match_all($pattern, $expr, $tokens);
print_r($tokens[0]);

?>

Using the heredoc syntax is not needed or recommended here.

answered Aug 1, 2011 at 15:06

ridgerunner

34.6k6 gold badges60 silver badges70 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Lea Hayes Over a year ago

Wasn't aware of that, thanks I will use that a lot. Is free-spacing mode specific to PHP or fairly standard?

ridgerunner Over a year ago

@Lea Hayes: Yes, most (PHP, Java, Perl, Python, .NET) have the 'x' mode - but unfortunately, JavaScript does not.

Lea Hayes · Accepted Answer · 2011-08-01 15:04:28Z

1

Finally, got there. It was all in the escape sequences of $pattern. Didn't work with HEREDOC or 'this type of literal' but did work in "this type of literal":

$pattern = "/\"(\\\\\"|[^\"])*\"|'(\\\\'|[^'])*'|[A-Za-z_][A-Za-z_\\-\\d]*|[\\-\\+]?\\d+(\\.\\d+)?|[=<>!][=]|[?:,()*\\/\\-\\+!]|\\|\\|?|\\&\\&?/";

I would be curious to see a working HEREDOC version though as I would prefer to avoid that nasty level of escaping within the pattern itself.

answered Aug 1, 2011 at 15:04

Lea Hayes

64.5k16 gold badges69 silver badges119 bronze badges

Collectives™ on Stack Overflow

Regex not working in PHP same as with C#

2 Answers 2

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related