Hi everyone!

I have a regex problem I could use your help with:

I am using PHP to preg_replace words within a block of text with hyperlinks, but when the word is already in an hyperlink, it is still turned into a link. Basically the preg_replace regex I am using is "double linking" the word.

In order to solve this I need some help in coming up with a regex that will match only words that are NOT in any html tags.

The preg_replace I am using now is a simple word match: preg_replace("/\bword\b/","<a href=\"http://www.example.com\">Word</a>",$text);

Thanks in advance for any help!

Craig

Please read the forum posting guide lines:

http://regexadvice.com/forums/thread/47451.aspx

Your question is, IMO, not clear.

This is a common question, consider this example:

http://regexadvice.com/forums/thread/45345.aspx

Hi

Thanks for the replies.

I have come up with the following (only a slight change to your previously posted regex :)

• (?s)\bFIND ME\b(?=([^<>]*<[^>]+>[^<>]*)+$|[^<>]*$)(?=((?!<body).)*[^?!(</body>|</a>)])

This seems to be working fine at finding all words within an html page NOT contained within tags

My next question is this - can you see anything wrong with the regex posted above? I have tested it using a few different scenarios and it seems to be fine, but I was just wondering if any of you regex guru's can see something I could be overlooking?

Thanks in advance,

Craig
 

This part: [^?!(</body>|</a>)] does not do what you think it does.

Everything between square brakets (called a character class) will match a single character and almost all of the "normal" meta characters have no special meaning inside character classes. So, [^?!(</body>|</a>)] will match any character except one of the following: ?, !, (, <, /, b, o, d, y, >, |, a, )

See: http://www.regular-expressions.info/charclass.html

Wistow:

I have come up with the following (only a slight change to your previously posted regex :)

• (?s)\bFIND ME\b(?=([^<>]*<[^>]+>[^<>]*)+$|[^<>]*$)(?=((?!<body).)*[^?!(</body>|</a>)])

As noted, that is a problematic change.

From the previous thread, this example might be of more help to your goal here:

(?si)\bFIND ME\b(?![^<>]*</a>)(?=([^<>]*<[^>]+>[^<>]*)+$|[^<>]*$)(?=(?:(?!<body).)*</body>)

See it in action here:

http://www.myregextester.com/?r=02f3271f

Thanks for the reply

If I were to change:

• (?s)\bFIND ME\b(?=([^<>]*<[^>]+>[^<>]*)+$|[^<>]*$)(?=((?!<body).)*[^?!(</body>|</a>)])

to:

• (?s)\bFIND ME\b(?=([^<>]*<[^>]+>[^<>]*)+$|[^<>]*$)(?=((?!<body).)*[^<])

It would essentially be doing the same thing? And how would I then add to the regex to find a string of characters e.g. </a> rather than a single character in the character class?

Thanks in advance,

Craig
 

ddrudik:

(?si)\bFIND ME\b(?![^<>]*</a>)(?=([^<>]*<[^>]+>[^<>]*)+$|[^<>]*$)(?=(?:(?!<body).)*</body>)

ddrudik, that seems to be exactly what I was looking for, and I now have more understanding of where I was going wrong with the regex I posted. (I posted my last reply unaware that you had replied, that's why it probably makes no sense :)

Thanks for all the help!

ddrudik:

(?si)\bFIND ME\b(?![^<>]*</a>)(?=([^<>]*<[^>]+>[^<>]*)+$|[^<>]*$)(?=(?:(?!<body).)*</body>)

With the above regex in mind, if I were to add |</span> after *</a>, so the full regex was:

• (?si)\bFIND ME\b(?![^<>]*</a>|</span>)(?=([^<>]*<[^>]+>[^<>]*)+$|[^<>]*$)(?=(?:(?!<body).)*</body>)

would this also mean the regex would exclude any text found in <span></span> tags? I believe it would, but i'm after an expert opinion :)

Thanks in advance!

Craig
 

It's best to test to see how a pattern would work with any given source, see your new pattern in action with source that contains a span:

http://www.myregextester.com/?r=c77b2c05

Also acceptable for your new task:

(?si)\bFIND ME\b(?![^<>]*</(?:a|span)>)(?=([^<>]*<[^>]+>[^<>]*)+$|[^<>]*$)(?=(?:(?!<body).)*</body>)