Regex to Extract between 2 delimiters

Anon

Hello,

I think this is a pretty common problem, for example, if we want to extract from a string with HTML code, all the strings that are between 2 certain tags.

Let's say:

$text = "blah blahthis is what I want 1yada yadathis is waht I want 2 and so on..."

And I want to extract the bold things, strings between $start="" and $end="".

We can suppose there are no nested tags.

I've searched in the site and I've found similar problems, but not the solution to this.

I can do a eregi with "$start(.*)$end" to get rid of everything that is before and after the first and , but I can't figure out how to get all the substrings I want.

I'm using PHP3, so I cannot use the Perl-like regexps available in PHP4.

Of course, I can do a brute-force approach and produce the worst algorithm using many explodes or splits:

function LazyExtract($start,$end,$text) {

$ExpStart = explode($start,$text);

for ($i=0 ; $i < sizeof($ExpStart) ; $i++) {
$ExpEnd = explode($end,$ExpStart[$i]);
for ($j=1 ; $j < sizeof($ExpEnd); $j++) {
$Result[$i] = $ExpEnd[0];
}
}
return $Result;
}

That code is terrible, I'm using a nested loop that will grow as a square.

Any help appreciated.

Anon

Sorry, I didn't know that the bold tags (b) and (/b) would actually format the message, but I guess it can be understood.

Anon

one thing to watch out for is the greediness of the ".*". there's not really an easy way around this without using perl..

one tactic relies on not having any other tags in the string; "$start([^<]*)$end" but obviously this may not suit you.

another could be to try and use split("</?b>",$text), and print out all the even numbered array elements.. this of course relies on the html being well formed.

--
rad

Anon

Thanks.

I wasn't sure if I could it without Perl.

The split() solution may be faster than the explode() one I mentioned, because it would use only one loop, although split() must be slower than explode().

Now... to look for a new PHP4 host service...

Anon

yes it can be done easely.

i sugest you break your string with all the html code in every , using explode

you can do something like this:

$cont = explode(" ",$HtmlString);

for($i = 0; $i < sizeof($cont); $i++)
{
$search=eregi("(.*)",$cont[$i], $write);

if($search)
{
print "write[1] "; //every line that was in bold
// the code when the criteria is found
....
}
}

Anon

sorry. the eregi function is delimited by the
bold tags. the rest is right