XPath：帮助在使用HTMLUnit抓取的DOM中查找特定元素

Question

I am scraping a webpage using HTMLUnit and have collected a List of DOM nodes from the webpage.

Inside each of these "company" DOM nodes is some data I want to scrape. For example I want the telephone number text from inside this node:

Now, this element would be a child of a div element which is in turn a child of another div element inside the company node. What would be the correct XPath line to access it? Here is my latest attempt which returned nothing.

 List<DomNode> companies = (List<DomNode>) page.getByXPath("//li[@class='featured block twoblock    boxshadow']");
        for (int j = 0; j < companies.size(); j++) {

            DomNode company = companies.get(j);

                // retrieve telephone number
                DomNode telephone = (DomNode) company.getByXPath(
                        "//li[@data-pvd-p='"+j+1+"']/div[@class='listingWrapper']/div[@class='itemInfo']/span[@class='tel']").get(0);

}

Here is a sample of the HTML, what:

        <li class="featured block twoblock boxshadow" data-pvd-p="3" data-pvd-c="0046176330000011028" data-pvd-et="sv" data-pvd-l="true">

    <div class="listingWrapper" itemtype="http://schema.org/LocalBusiness" itemscope="">
        <a href="/Craddock-Electrical-Services-Ltd/0046176330000011028/"></a>
        <div class="itemInfo">
            <div class="tradeImage" itemprop="member" itemscope="" itemtype="http://schema.org/Organization"></div>
            <h2>
                <a itemprop="name" href="/Craddock-Electrical-Services-Ltd/0046176330000011028/"></a>
            </h2>
            <span class="tel" itemprop="telephone"></span>
            <div class="listLinks"></div>
            <div id="addressBar"></div>
        </div>
        <div class="itemInfo2"></div>
        <div class="clearLeft"></div>
        <ul class="features"></ul>
        <div class="clearLeft"></div>
        <p class="promo" itemprop="description"></p>
    </div>
</li>

UPDATE 2:

Here is the current state of my XPath code.

List<DomNode> companies = (List<DomNode>) page
                .getByXPath("//li[contains(@class, 'featured block')]");
        for (int j = 0; j < companies.size(); j++) {

            String url = "";
            DomNode company = companies.get(j);
            DomElement web = null;

            // retrieve name
            DomNode name = (DomNode) company.getByXPath("//a[@itemprop='name']").get(j);

            if (companiesLogged.contains(name.getTextContent().trim()) != true) {
                companiesLogged.add(name.getTextContent().trim());

                // retrieve telephone number
                DomNode telephone = (DomNode) company.getByXPath("div[@class='listingWrapper']/div[@class='itemInfo']/span[@class='tel']").get(0);


                // retrieve website
                try{
                web = (DomElement) company.getByXPath("div[@class='listingWrapper']/div[@class='itemInfo']" +
                        "/div[@class='listLinks']/a[@id='linkWebsite']").get(0);
                } catch(IndexOutOfBoundsException e){
                    System.out.print(" (No Website) ");
                }

                try{
                url = web.getAttribute("href");
                } catch (IndexOutOfBoundsException e){
                    url = "N/A";
                }

                System.out.println(name.getTextContent().trim() + "   "
                        + telephone.getTextContent().trim()
                 +"   "+url.trim());

            } else {
                System.out.println("Company already logged");
            }
        }

Answer 1

First thing I see is how you're retrieving the group of <li> nodes. Just looking at your @class attribute, you can't really tell how many spaces are in " featured block twoblock boxshadow ", but that XPath will only return a result if it is exactly equal to it . In that regard, try using something more flexible like contains() , ie //li[contains(@class, 'featured block')] .

Without seeing what source you're targeting I can't suggest much more, but will update the answer when it's added to the question.

I've tried your XPath (just the /div part, since that's what was provided) on the given snippet and got back <span class="tel" itemprop="telephone"/> as a result. Looks like an issue with how you're retrieving the <li> company nodes.

Update 2: From your updated XML snippet, your first XPath //li[@class='featured block twoblock boxshadow']" doesn't look like it will match the parent <li> node, based on what I mentioned with the spaces before. Secondly if it did, you are checking the <li> node's attributes twice on separate queries, and assuming that the index you're giving the data-pvd-p value (starts at 3 in the snippet) will always match the list index (starts at 0, with your +1 added). I'd suggest removing this portion //li[@data-pvd-p='"+j+1+"'] and beginning with the //div .

So something like this:

List<DomNode> companies = (List<DomNode>) page.getByXPath("//li[contains(@class, 'featured block']");
        for (DomNode node : companies) {

                // retrieve telephone number
                DomNode telephone = (DomNode) node.getByXPath(
                        "div[@class='listingWrapper']/div[@class='itemInfo']/span[@class='tel']").get(0);