Regular Expression Question

imi_99

I want to search and replace a string that is not in
the span tag. Basically ignore tags and carry out the
susbitution.

$searchWord ="background";
$replaceString = "";
$text = "A <span
style=background-color:#00ff00>Green</span>background
colour";

when I use preg_search function to replace background
word with $replaceString it subsitutes background in
span tag.

Can some one help me please, so I can I avoid
susbsituting span tag content?

Thanks

rocky2083

Why don't you use ' or escape \" when you start styles and this way you can partly delimit style backgrounds with other text. Rewrite expressions to suit it.

rincewind456

Don't know if this will work for you, but if background is a word in text then it will most likely have a space at the right hand end of it so if you use that space in the search string it wont match the background in background-color. like this:

<?php
$searchWord ="/background /";
$replaceString = " ";
$text = "A <span style=background-color:#00ff00>Green</span>background colour";

echo preg_replace($searchWord,$replaceString,$text);
?>

madwormer2

Or you could do more than one replace...

First replace background-color with something like {SPAN_BG_COLOR} (or at least something that is very unlikely to be a part of the text)
Then do the replace on the word "background"
And then, do another replace on {SPAN_BG_COLOR} back to background-color

It can actually be achieved like:

str_replace( Array('background-color:','background','{CSSBGCOLOR}') , Array('{CSSBGCOLOR}','','background-color:'), $text );

Maybe... I'm not sure if that was what you were after.

xblue

Maybe the lookahead example here comes close to what you want.

(It would not work out on your $text since it requires the whole text be wrapped in a tag (like <html> in a complete html document), but should not be to difficult to modify.)

imi_99

xblue wrote:
Maybe the lookahead example here comes close to what you want.

(It would not work out on your $text since it requires the whole text be wrapped in a tag (like <html> in a complete html document), but should not be to difficult to modify.)

Thanks xblue, I got somewhere with this "tote-taste.de" regex info. but i have a minor issue.

What I did is appended the text within <html>$text</html>. Its pretty useless having tag around all my search text.
As you have mentioned, its simple to remove the begining and ending tag with the given regex. Can you please shed some light on it. No point me replacing it with another call to str replace function if I cna do it with in the regex.

I used positive lookaheads.

Eg:
$text = "<html>$text</html>";

preg_replace("/($term)(?=[^>]*<)/ise" ....

How can I have the above regex so it does not require all the text to be in tags?

Many thanks.

xblue

Ok, let's take the lookahead
(?=[^>]*<)

This represents the part "our keyword is followed by an opening angular bracket without a closing angular bracket in between".

What modification you need is ""our keyword is followed either by an opening angular bracket or by the end of the string without a closing angular bracket in between".

So, you need to replace the opening angular bracket in the above with the new condition
(?=[^{>]*COLOR=DarkRed[/COLOR])}
It is not tested, but I am sure you get the logic behind that.

imi_99

Thanks to all, who shared their thoughts and corrected me.

But I must say, xblue is a star - explained with details - Many tahnks

ckryzan

I am having a (really) hard time crafting a regular expression to achieve what seemed to be relatively straightforward task: splitting a string into two pieces using a period as the delimiter, subject to some conditions (specifically to NOT split URLs, email addresses or real numbers).

I want to split substrings of the form "before abc.def after", "before a12.b3c after", extracting the components from the larger string (yielding "abc", "def" and "a12" and "b3c", for example. The substrings I will be testing are uniform in that they are to contain only a single period, and the portions I want to extract are bounded by whitespace or start/end characters (although probably any "\W" would also be allright.

I do NOT want strings of the following types to be extracted/split: "before email123@host.com after", "before email@host123.com after", "email.more@subhost.host.com", "server.host.domain", "123.456".

My PHP code to extract these substrings is:

$string = preg_replace("/(?=([\s^{][a-zA-Z_]\w*)\.\w+[\s$])(?=\1)\./","} ",$string);

The idea behind this regex is to test the string for conformance, then to move the pointer up to the decimal point and use that as the split trigger point.

After hours and hours and hours of tweaking, I have been unable to get this or anything close to it to work.

I would very much appreciate if anyone could offer insight into what I might be doing wrong. Thanks!

Chris