[RESOLVED] Parsing Text from text file between code

BASiQ

I have a text file with code like this

<th>Name</th>                        <th>Item</th>                        <th>Item Id</th>                </tr>        </thead>        <tbody><tr class="tableListLeft">
<td><div style="float: left; margin: 0pt 6px 12px 0pt;" class="iconHolder_in-game"><div class="avatarIcon"><a href="/profiles/76561197987317160/"><img alt="Master Blaster" src="http://media.steampowered.com/steamcommunity/public/images/avatars/a1/a168e4b4336b505ee1f4bf460348c40a81467279.jpg"/></a></div></div><span style="float: left; font-size: large; line-height: 20px;"><a href="/profiles/76561197987317160/">Master Blaster</a></span><br/><span class="friendBlock_in-game">In-Game: Team Fortress 2</span></td>
<td style="text-align: center;"><div class="invBoxSmall"><img alt="Unusual Pugilist's Protector" src="http://mirror.pointysoftware.net/tf2items/hats-heavy/pugilist_protector_sized.png" class="item" id="item1"/>
</div></td><td style="text-align: right;"><a href="/item/688749855">688,749,855</a></td>

</tr>
<tr class="tableListAlternateLeft">
<td><div style="float: left; margin: 0pt 6px 12px 0pt;" class="iconHolder_in-game"><div class="avatarIcon"><a href="/profiles/76561198037581884/"><img alt="Azbest Sell items" src="http://media.steampowered.com/steamcommunity/public/images/avatars/36/36b041f262deeba26ca7e197ac1bf578d42f6624.jpg"/></a></div></div><span style="float: left; font-size: large; line-height: 20px;"><a href="/profiles/76561198037581884/">Azbest Sell items</a></span><br/><span class="friendBlock_in-game">In-Game: Team Fortress 2</span></td>
<td style="text-align: center;"><div class="invBoxSmall"><img alt="Unusual Pugilist's Protector" src="http://mirror.pointysoftware.net/tf2items/hats-heavy/pugilist_protector_sized.png" class="item" id="item2"/>
</div></td><td style="text-align: right;"><a href="/item/688583590">688,583,590</a></td>

I need to use a php file to call the text file,
find the lines like this /profiles/76561198037581884/
and just grab the numbers from the end and echo them to another txt file or within the page, either is fine.

This should be super simple. I know how to call the text file.
I have and use simple dom for parsing. I just dont know how to parse text inbetween text.

Thanks! =)

sneakyimp

// I assume that the var $html contains the html in your previous post:
// this script will output the numbers into this file.  we will replace its contents with the new results
$output_file = "output.txt";
$pattern = '#<a[^>]+href="/profiles/([0-9]+)/"#i';
$matches = NULL;
$match_count = preg_match_all($pattern, $html, $matches);
if ($match_count > 0) {
  echo $match_count . " matches found\n";
  if (!file_put_contents($output_file, implode($matches[1], "\n"))) {
    die("Could not write $output_file\n";
  }
  echo "success\n!";
} else {
  die("No matches found!");
}

bradgrafelman

If you're just interested in scanning the entire text and finding all matches based on the pattern of '/profiles/<numbers>/', then I'd say DOM is irrelevant - just use something like [man]preg_match_all/man with a pattern like:

@/profiles/([0-9]+)/@

EDIT: Woops, guess I should refresh old tabs before replying.

BASiQ

Got it working

$filein = file('TF2U.txt');
foreach ($filein as $html) {
	$pattern = '#.*<a[^>]+href="/profiles/([0-9]+)/"#iA';
	$matches = NULL;
	$match_count = preg_match_all($pattern, $html, $matches);
	if ($match_count > 0) {
		echo implode($matches[1]);
		echo "<br>\n";
		}

Decided to skip the output to a text file as it was making my work slow.
Thanks for the help guys! I now have a perfectly functioning script lol

johanafm

bradgrafelman;10995646 wrote:
If you're just interested in scanning the entire text and finding all matches based on the pattern of '/profiles/<numbers>/', then I'd say DOM is irrelevant

Personally, I still like using DOM to exclude anything irrelevant. In this case it's for example easy to disregard multiple occurances of /profiles/12345/, although I admit that it seems that they are likely to always occur the same number of times (2) and that it can be handled by regexp.

I also find it to be a lot more clear on exactly what is beeing looked for in the html code, assuming the full path is explicitly stated.

$d = new DOMDocument();
$d->loadHTML($filein);
$x = new DOMXPath($d);

# Find only a element whose href starts with '/profiles/' in this particular place in the DOM tree
$l = $x->query("/html/body/table/tbody/tr/td/div/div/a[starts-with(@href,'/profiles/')]");

# Or to make the above equivalent of the reg exp (which finds such an a element anywhere in the tree)
# $l = $x->query("//a[starts-with(@href,'/profiles/')]");


$len = strlen('/profiles/');
foreach ($l as $n)
{
	$h = $n->getAttribute('href');
	# or write to file
	echo substr($h, $len, -1) . PHP_EOL;
}

bradgrafelman

johanafm;10995662 wrote:
Personally, I still like using DOM to exclude anything irrelevant.

I agree with this and pretty much everything that followed in your reply, which is why I prefaced my answer with an 'If'.

Based on the OP's request:

BASiQ wrote:
find the lines like this /profiles/76561198037581884/

it sounded like the interest wasn't in treating the data as HTML and looking for a node/element (or something higher level like a "hyperlink"). Instead, it seemed like the intent was to treat the data as nothing more than plain text in a file and to locate instances of a pattern within that blob of text.

However, based on the appearance of the '<a href' syntax in the OP's subsequent reply, I would say this is yet another case of someone not saying what they mean (or not meaning what they say) and, as such, my reply is no longer relevant. :p