XPath在TR中选择TD

I want to capture all the content between td tags but divide them by their tr. So i can get an array with the content inside every tr.

<div id="box">
<tr align='center'>
<td>1</td>
<td style='padding-left: 0px !important;padding-right: 10px !important;'>   <div id=''></div></td> 
<td>45</td>
<td>62</td>
</tr><tr align='center'>
<td>2</td>
<td style='padding-left: 0px !important;padding-right: 10px !important;'>   <div id=''></div></td> 
<td>35</td>
<td>47</td>
</tr><tr align='center'>
<td>3</td>
<td style='padding-left: 0px !important;padding-right: 10px !important;'>   <div id=''></div></td> 
<td>63</td>
<td>58</td>
</tr>

I've tried with this:

<?php
$url = '';
$html = file_get_contents($url);
$doc = new DOMDocument();
$doc->preserveWhiteSpace = FALSE;
@$doc->loadHTML($html);
$xpath = new DOMXpath ($doc);
$expresion = "//div[@id='box']//tr//td";
$node = $xpath->evaluate($expresion);
foreach ($node as $nd)
{
echo $nd->nodeValue;
}
?>

But the output is:

1

45
62
2

35
47
3

63
58

If you want to group the td values by their tr, I would separate the xpath into two queries. One query selects the <tr> nodes and a second query selects the <td> childs of that node.

If you put that into a loop it can look like this:

<?php

$html = <<<EOF
<div id="box">

    ... Your HTML comes here
</tr>
EOF;

$url = '';
$doc = new DOMDocument();
$doc->preserveWhiteSpace = FALSE;
@$doc->loadHTML($html);
$xpath = new DOMXpath ($doc);
$expresion = "//div[@id='box']//tr";
$trs = $xpath->evaluate($expresion);
foreach ($trs as $tr)
{
    $tdvals = array();
    foreach($xpath->query('td', $tr) as $td) {
        /* Skip the td with the empty text value */
        if(trim($td->nodeValue) !== '') {
            $tdvals []= $td->nodeValue;
        }
    }
    echo implode(',', $tdvals) . PHP_EOL;
}

which outputs:

1,45,62
2,35,47
3,63,58

One another thing. In your example you are using file_get_contents() to load the HTML. Note that you can use DOMDocument::loadHTMLFile() to load (remote) files.