[RESOLVED] (Probably) simple regexp messing me up

bunner_bob

I know this is easy, but I'm currently vexed...

I'm trying to get rid of any tags that occur before list tags (ul, ol, li, etc.), but it's getting rid of other tags I want to keep.

The code looks like:

$cleaned= preg_replace('/<br \/>([<ul>|<UL>|<\/ul>|<\/UL>|<ol>|<OL>|<\/ol>|<\/OL>|<li>|<LI>|<\/li>|<\/LI>])/', '\\1', $text);

An excerpt of source text might be:

<br /><li><strong>Customer Referrals</strong><br />Members information will be posted for clients on a page specific to your state and location.... to the general public.</li><br /><li><strong>Website Listing</strong><br />One of the most important ways ... database of qualified professionals.</li>

I'm trying to get rid of the tags preceding the <li> and </li> tags, and leave alone the ones that follow the tags above. The first one (Customer Referrals) is fine - the is left alone. But in the second (Website Listings) the tag following is replaced - not what I want.

The result looks like this:

<li><strong>Customer Referrals</strong><br />Members information will be posted for clients on a page specific to your state and location.... to the general public.</li><li><strong>Website Listing</strong>One of the most important ways ... database of qualified professionals.</li>

Naturally this is driving me batty, but I bet I'm just doing something stupid. Can anyone tell me what's wrong?

whisher06

Hi.
Try this:

<?php
$string = "<br /><br><li><strong>Customer Referrals</strong><br />Members information will be posted for clients on a page specific to your state and location.... to the general public.</li><br /><li><strong>Website Listing</strong><br />One of the most important ways ... database of qualified professionals.</li> ";
$array = array("<br />","<br>");
$cleanerBr = str_replace($array,"",$string);
echo  $cleanerBr;
?>

Bye 🙂

Houdini

Actually Whisher's code works but it eliminates all or which in the case of a list item doesn't matter because when a new list item happens it starts it own new line without the or or \r\n or \n. So if you ran his code and looked at the source you would see that all the or have been removed, but it still gives the intended result.

xblue

The problem with your pattern are the square brackets. The turn what was meant to mean "either this or this or that tag" into a character class of [ouOU|\/<>] and the match the O in One.

'/<br \/>(<ul>|<UL>|<\/ul>|<\/UL>|<ol>|<OL>|<\/ol>|<\/OL>|<li>|<LI>|<\/li>|<\/LI>)/'

should work more the way you want.

Apart from that, a few suggestions:
- Use a different delimiter, e.g. #, then you don't need to escape the slashes in your pattern.
- Since the angular brackets repeat, the can be placed outside the alternation.
- Similar for the slash, as you want each tag either with or without, you can make it optional don't need to repeat all the tags.
- the i modifier saves you from covering all capitalization possibilities.

The pattern rewritten:

'#<br />(</?(ul|ol|li)>)#i'

bunner_bob

xblue wrote:
Apart from that, a few suggestions:
- Use a different delimiter, e.g. #, then you don't need to escape the slashes in your pattern.
- Since the angular brackets repeat, the can be placed outside the alternation.
- Similar for the slash, as you want each tag either with or without, you can make it optional don't need to repeat all the tags.
- the i modifier saves you from covering all capitalization possibilities.

Thank you thank you thank you!

I'm having one of those moments when I feel "lifted up" by a rush of understanding. I thought I had a passable understanding of regular expressions, but you've just kicked me up a good notch.

Clearly I was writing my original expression in "lazy mode". I'm going to have to revisit all my other expressions now.

I didn't know you could use a different delimiter, and I never thought about the case insensitive modifier. I must go re-read all my regexp links now...

btw - Whisher's code wouldn't have worked. I was trying to retain tags that weren't adjacent to list item tags. It was just the list item ones I was trying to kill - they make the spacing of the list items get all goofy.

whisher06

:o Just a red face !
Bye. 🙂

xblue

No problem, glad to be of service 😉

As you appreciated my answer (I think I tend to be too complicated sometimes, but I am no native English speaker which may serve as partial excuse), maybe you enjoy something I put together on regular expressions.