使用XPATH节点从产品获取数据并将其插入表中

Question

I'm trying to get product data from the external web site and insert them in special table - every found node element need to be imported in appropriate column for the product in the product table!

It's working fine for finding 1 product attribute and inserting it into table:

$product_names = $xpath->query("//div[contains(concat(' ', normalize-space(@class), ' '), ' product_description ')]/div/h3/a");
        if (!is_null($product_names)) {
            foreach ($product_names as $product_name) {
                $nodes = $product_name->childNodes;
                foreach ($nodes as $node) {
                    $import_product = 'INSERT INTO product_table (id, product_name) values ("","' . preg_replace('~\\s+\\S+$~', "", strip_tags(trim($node->nodeValue))) . '")';
                    mysql_query($import_supralift_name);
                }
            }
        }

but products have many attributes, so, I'm trying to get this product attribute (which is in 1 html element, so I need to split it in array an use for different attributes):

$types = $xpath->query("//div[contains(concat(' ', normalize-space(@class), ' '), ' product_description ')]/div/a/p");
        if (!is_null($types)) {
            foreach ($types as $type) {
                $nodes = $type->childNodes;
                foreach ($nodes as $node) {
                    list($typee,$power_unit) = explode(' / ', $node->nodeValue);
                    $import_type = 'INSERT INTO product_table (id, type, power_unit) values ("", "' . strip_tags(trim($typee)) . '", "' . strip_tags(trim($power_unit)) . '")';
                    mysql_query($import_type);
                }
            }
        }

In short - I need to get 3 product attributes (of course, they're more, just want to figure out what would be the best solution to get it working) from external web site and insert it into my data base like:

product_name_1 product_type_1 $power_unit_1
...
product_name_X product_type_X $power_unit_X

So far I tried to put second xpath part in first foreach, but it does not work as need... Should I try to make array with xpath nodes (like $prodcuts=array(firstXpathNode, secondXpathNode etc..) and work in such way or there is better and more correct solution for this?

In advance - TXN for any tips...

EDITED: Here are sample HTML from what I'm trying to get the data, this is for the product (each product has this html for displaying data):

<div class="single_product">
    <div data-section="featured_image">
        <a title="Unique_String" href="#">
            <div style="" data-section="image" class="image_in_fixed_ratio_wrapper">
                <div class="inner visible">
                    <img alt="Unique_String" src="image1.jpg" class="" style="">
                </div>
            </div>
        </a>
    </div>
    <div data-section="data">
        <div class="product_description">
            <div data-field="description_detail">
                <h3><a title="Unique_String" href="#">Product Name<div class="donotwantthistoinclude">New</div></a></h3>
                <a title="Unique_String" href="#"><p>Product Type / Product Power Unit</p></a>
                <div data-field="price">
                    <a title="Unique_String" href="#">5,000</a>
                </div>
                <div data-field="description">
                    <a title="Unique_String" href="#">
                        <span>Height (mm)</span> 2344
                 |
                                <span>Other attribute 1</span> Duplex
                 |
                                <span>Other attribute 2 (kg)</span>  1400
                 |
                                <span>Other attribute 3</span> 2014

                                 | <span>Other attribute X (h)</span> 772
                        <br><span>Location</span> D - 85716
                    </a>
                </div>
            </div>
        </div>
    </div>
</div>

Answer 1

If you separate out the product name in the first foreach into a variable, you can build a relative XPATH based on the product name. I'm assuming the product names are unique on the page. Then the second XPATH finds the product name on the page and walks a bit further down the elements. Now, there will guaranteed be better XPATH queries to write to do that, I just haven't got that skill level myself, but I do give you one way to do it.

The flow will therefore be something like:

for each product, get the name, insert name into new query to get that particular product's type and power unit, parse the variables, insert into DB.

WARNING

You are using dangerous and outdated SQL. Please use the newer mysqli_* or PDO libraries to access the database using prepared statements. I did NOT update your code to reflect that, it's easy to Google.

I did however insert product_name in your existing SQL to illustrate how all 3 fields are gathered.

$product_names = $xpath->query("//div[contains(concat(' ', normalize-space(@class), ' '), ' product_description ')]/div/h3/a");
if (!is_null($product_names)) {
    foreach ($product_names as $product_name) {
        $nodes = $product_name->childNodes;
        foreach ($nodes as $node) {
            $productName = preg_replace('~\\s+\\S+$~', "", strip_tags(trim($node->nodeValue)));
            $xpath_relative = sprintf("//div[contains(concat(' ', normalize-space(@class), ' '), ' product_description ')]/div/h3/a[contains(text(),'%s')]/../../a/p",$productName);

            $types = $xpath->query($xpath_relative);
            if (!is_null($types)) {
                foreach ($types as $type) {
                    $types_nodes = $type->childNodes;
                    foreach ($types_nodes as $type_node) {
                        list($typee,$power_unit) = explode(' \'', $type_node->nodeValue);

                        // WARNING!!! SQL INJECTION BELOW!!!
                        $import_type = 'INSERT INTO product_table (id, type, power_unit, product_name) values ("", "' . strip_tags(trim($typee)) . '", "' . strip_tags(trim($power_unit)) . '", "' . $product_name . '")';
                        mysql_query($import_type);
                    }
                }
            }
        }
    }
}

EDIT #2

I have taken your code and run with it in a PHP Fiddle with the following result. I've also optimised the XPATH queries based on the provided structure, as well as provide a suggestion on using PDO. Just fill in more attributes as needed. I'll leave you with the entire code, including the DOM and XPATH initialisation I have used so you can fiddle with it yourself.

<pre><?php

$domDoc = <<<EOF
<div class="single_product">
    <div data-section="featured_image">
        <a title="Unique_String" href="#">
            <div style="" data-section="image" class="image_in_fixed_ratio_wrapper">
                <div class="inner visible">
                    <img alt="Unique_String" src="image1.jpg" class="" style="" />
                </div>
            </div>
        </a>
    </div>
    <div data-section="data">
        <div class="product_description">
            <div data-field="description_detail">
                <h3><a title="Unique_String" href="#">Product Name<div class="donotwantthistoinclude">New</div></a></h3>
                <a title="Unique_String" href="#"><p>Product Type / Product Power Unit</p></a>
                <div data-field="price">
                    <a title="Unique_String" href="#">5,000</a>
                </div>
                <div data-field="description">
                    <a title="Unique_String" href="#">
                        <span>Height (mm)</span> 2344
                 |
                                <span>Other attribute 1</span> Duplex
                 |
                                <span>Other attribute 2 (kg)</span>  1400
                 |
                                <span>Other attribute 3</span> 2014

                                 | <span>Other attribute X (h)</span> 772
                        <br /><span>Location</span> D - 85716
                    </a>
                </div>
            </div>
        </div>
    </div>
</div>
EOF;
$dom = new DomDocument();
$dom->loadXML($domDoc);
$xpath = new DomXPath($dom);

$products = [];

$productUniqueQuery = "//div[@data-field='description_detail']/h3/a/@title";

$productUniqueNodes = $xpath->query($productUniqueQuery);
if (!is_null($productUniqueNodes)) {
    foreach ($productUniqueNodes as $productUniqueNode) {
        $product = [];
        $product["unique"] = $productUniqueNode->nodeValue;

        $productNameQuery = sprintf("//h3/a[@title='%s']/text()",$product["unique"]);
        $productNameNodes = $xpath->query($productNameQuery);
        $product["name"] = $productNameNodes[0]->nodeValue;

        $productImageQuery = sprintf("//img[@alt='%s']/@src",$product["unique"]);
        $productImageNodes = $xpath->query($productImageQuery);
        $product["imageURL"] = $productImageNodes[0]->nodeValue;

        $productTypeQuery = sprintf("//a[@title='%s']/p/text()",$product["unique"]);
        $productTypeNodes = $xpath->query($productTypeQuery);
        list($product["type"], $product["powerUnit"]) = explode(" / ", $productTypeNodes[0]->nodeValue);

        $productDescriptionQuery = sprintf("//div[@data-field='description']/a[@title='%s']/child::node()",$product["unique"]);
        $productDescriptionNodes = $xpath->query($productDescriptionQuery);
        $description = "";
        foreach ($productDescriptionNodes as $productDescriptionNode) {
            $nodeText = preg_replace("/\s*\|/","",trim($productDescriptionNode->nodeValue));
            if($nodeText == "" || $productDescriptionNode->nodeType === 3){
                continue;
            }

            $product[$nodeText] = preg_replace("/\s*\|/","",trim($productDescriptionNode->nextSibling->nodeValue));
        }
        $products[$product["unique"]] = $product;
    }
}


try {
    $db = new PDO("mysql:host=HOST;dbname=DBNAME;port=3306","USERNAME", "PASSWORD");
}
catch(PDOException $e){
    echo "Connection failed: " . $e->getMessage();
    exit();
}

$sql = 'INSERT INTO product_table (unique, name, type, power_unit, attr1) values (:unique, :name, :type, :power_unit, :attr1)';
$stmt = $db->prepare($sql);

foreach($products as $product){
    $params = [
        ":unique"=>$product["unique"],
        ":name"=>$product["name"],
        ":type"=>$product["type"],
        ":power_unit"=>$product["powerUnit"],
        ":attr1"=>$product["Other attribute 1"]
    ];
    var_dump($product);
    $stmt->execute($params);
}

?>
</pre>

Answer 2

One thing you can do to make it easier, is when using XPath, you can use one node as the context of further searches, so once you have a list of product nodes, use this as a point at which you extract the other data.

Just as an example...

$dom = new DomDocument();
$dom->loadXML($xml);
$xpath = new DomXPath($dom);

$products = [];

$data = $xpath->query("//div[@class='single_product']");
foreach ($data as $item) {
$name = $xpath->evaluate('string(descendant::div[@data-field="description_detail"]/h3/a/@title)'
        ,$item);

$imageName =  $xpath->evaluate('string(descendant::div[@data-section="featured_image"]//img/@src)'
        ,$item);
$typePower = $xpath->evaluate('string(descendant::div[@data-field="description_detail"]/a/p/text())'
        ,$item);
$description = $xpath->evaluate('string(descendant::div[@data-field="description"]/a)'
        ,$item);

    $products[$name] = array( "image" => $imageName,
            "typePower" => $typePower,
            "description" => $description
    );
}

print_r($products);

Note the second parameter to the evaluate() method which is the node from the first query() .

I've also used evaluate which allows me to return the node as a string straight away without any further conversion (it allows me to use string() as part of the query).

There is no post processing, so you may have to tidy up some of the data and there is no database access (you should follow the examples of using prepared statements), but this shows the important part of extracting the data in the first place.