Why does my QueryPath script result in "Creating default object from empty value" error?

Question

I am trying to scrape a website to get some practice with QueryPath.

Here is what I have so far, and gives me an error:

Creating default object from empty value

Code:

// URL to scrape
$baseurl = 'http://some-site-with-a-table-of-items-that-contain-links.com';

// Get all rows from table
$rows = htmlqp($baseurl, '#items_table')->find('tr');

//initialize items array
$items = array();

// initilize counter
$i = 0;

// Iterate through rows of items
foreach($rows as $row) {

    // get the url for the item in this row
    $url = qp($row)->find('.link_txt a')->attr('href');

    // select all the info in the item detail box
    $item = htmlqp($url)->find('.item_detail_box');

    // assign the item attributes to an array
    $items[$i] = [

        // the qp item $row is from the info on the main table of items
        'img_thumb'     => qp($row)->find('.reflection')->attr('src'),
        'name'          => qp($row)->find('.link_txt a')->text(),
        'item_level'    => qp($row)->find('.col_center')->text(),
        'req_level'     => qp($row)->find('.col_right')->text(),
        'url'           => $url,

        // the qp item $item is from the actual item detail page
        //'img'         => qp($item)->find('.reflection')->attr('src'),
        //'is_unique'   => qp($item)->find('.unique')->text(),


    ];

    $i++;
}

$data = print_r($items, true);

return '<pre>' . $data . '</pre>';

The error will occur if I uncomment either of the img or is_unique array lines.

Everything else works and gives expected output when those lines are commented out.

Answer 1

The problem happened because QueryPath was getting nothing from the selector trying to get text from an anchor tag.

I was trying to get the text from a link/anchor from each table row.

However, the first row in my loop was a table header and not a row with any links.

Adding a check in the loop fixed my issue:

$url_ext = qp($row)->find('.ic_link_txt a')->attr('href');

    if ( $url_ext != NULL && $url_ext != "" ) {

Which was a stupid mistake on my part not knowing enough about QueryPath.

(Also related to github issue https://github.com/technosophos/querypath/issues/130 )