Jsoup Tutorial: Parsing a website with Java

Jsoup is a library allowing us to

  • parse HTML from an URL, file or string
  • Extract data from the DOM, i.e. by using CSS selectors
  • Manipulate HTML elements, attributes and texts
  • Generate clean HTML

In this Jsou tutorial, I am going to show you how to parse a website and extract data.

First off, you need to include the jsoup dependency. For news, changelog and bug reports, check out the official Jsoup page.


Parsing a document

Start by parsing HTML into a Document. Henceforth, we are working with that Document. Jsoup is called statically. Jsoup offers the following variants of parsing HTML.

Jsoup - Methoden zum Parsen eines Dokumentes
Jsoup – Methods for parsing HTML

To parse a website to a document, you can simply call the following method


import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;

public class WebsiteParser {

  private static final int TIMEOUT_IN_MS = 5000;
  public static void main(String[] args) throws MalformedURLException, IOException
    Document doc = Jsoup.parse(new URL("url"), TIMEOUT_IN_MS);

Extract Data

Once we parsed the HTML to a document, we can start extracting data. Jsoup offers quite a few ways to do so, for example…

  • getElementById – Returns a HTML element (org.jsoup.nodes.Element) by the given ID
  • getElementsByTag – Returns a list of HTML elements ( with the given tag
  • getElementsByClass – Returns a list of HTML elements ( with the given CSS class
  • getElementsByAttribute – Returns a list of HTML elements ( containing the given attribute
  • select – Returns a list of HTML elements ( matching the given CSS selector

You can also apply further selectors on, making the following example valid

Elements tables = doc.getElementsByAttribute("table");
for(Element table : tables)
   Elements tableRows = table.getElementsByAttribute("tr");
  for(Element tableRow : tableRows)
    Elements tableData = tableRow.getElementsByTag("td");
     // ...

CSS Selector

As mentioned previously, the select method makes it possible to select HTML elements by providing a CSS selector.

Elements links ="a[href]"); // Get all links with a href attribute present
Elements gifs ="img[src$=.gif]"); // Get all images that are gifs

Element tableRows ="table.versions > tbody > tr").first(); // Get all table rows "tr" within a tbody element, in a table element with the class versions

Elements resultLinks ="span > a"); // Get all links within a span element
Elements lastTableChild ="table > tbody > tr:last-child"); // Get the last child of a table row within a table body within a tables

Overview of selectors

  • tagname: Finds the element with the given tag, i.e. a or div
  • ns|tag: Finds the elements from a given tag in a specific namespace, fb|name matches <fb:name> elements
  • #id: Finds elements by the given ID, i.e. #eow-title
  • .class: Finds elements with the given class, i.e. .versions finds all elements with the versions class
  • [attribute]: Finds all elements with the given attribute, i.e. [href] or [name]
  • [^attr]: Finds all ements with the given prefix as attribute name, i.e. [^data-] findet alle elements with HTML5 data-attributes
  • : Finds all elements with the given attribute and value, i.e. [width=500]
  • , , : Finds all elements, that start/end/contain the attributes value, i.e. [href*=/path/] finds all elements, whose href attributes contain /path/ in it's value
  • : Finds all elements, matchign the given regex, i.e. img[src~=(?i)\.(png|jpe?g)]
  • *: Finds all elements

Combining selectors

  • el#id: Elements with the ID, i.e. div#logo
  • el.class: Elements with the class, i.e. div.masthead
  • el: Elements with the given attribute, i.e. a[href]
  • Concatted, i.e. a[href].highlight
  • ancestor child: Child elements, that have a specific parent, i.e. .body p finds all p elements, that are children (not just directly), of an element with a body class
  • parent > child: Finds the direct child element, i.e. div.content > p finds all p elements, that are direct children of a div with the class content
  • el, el, el: Group multiple selectors, i.e. div.masthead, div.logo

List of all selectors on

Real-World example – Parsing dependency versions from

Theory is one thing, but let us see Jsoup in action with a real-world example.

Let’s try parsing the dependency versions from a library, let’s take Async-Http-Client from com.ning.

This is how the source code of the page looks like:

<table class="grid versions" width="100%">
         <td rowspan="56">
         <td><a href="/artifact/com.ning/async-http-client/1.9.30" class="vbtn release">1.9.30</a></td>
            <div><a href="/artifact/com.ning/async-http-client/1.9.30/usages">2</a><span class="rankingbar" style="width: 7px; "></span></div>
         <td> (Jul, 2015) </td>

We want to parse the exact version (1.9.30), the type (release) – could be release, beta, general availability or something else and the usages (2).

Let’s create a POJO representing a dependency.

public class MavenDependency {

  private String group;
  private String artifact;
  private String version;

  private int usages;
  private String type;

 // Konstruktor

 // Getter & Setter

 // ToString


Start by parsing the website

// Parse website with a 5 secodn timeout
Document document = Jsoup.parse(new URL(""), 5000);

Using the css selector table.versions > tbody > tr, we get all rows containing the versions. It is important to select the tbody, else we wil also get unnecessary information from the tables’s head.

Elements versionRows ="table.versions > tbody > tr");

Now that we have the rows, we continue extracting the exact information for a single dependency entry by applying CSS selectors.

  • Version: td > a.vbtn – Link with the class vbtn in a td element
  • Usages: td:nth-last-child(3) – The third last child from the td element
  • Type: td:nth-last-child(2) – The second last child from the td element

There are probably a bunch of other ways to access this data. This is how the entire code looks like

public class MavenParser {

  private static final String MVN_REPO_BASE_SEARCH_URL = "";

  public static void main(String[] args) throws MalformedURLException, IOException {
    List<MavenDependency> availableDependencies = getMavenResults("com.ning", "async-http-client"); // ->

  public static List<MavenDependency> getMavenResults(String group, String artifact) 
      throws MalformedURLException, IOException {
    String url = MVN_REPO_BASE_SEARCH_URL + group + "/" + artifact;

    // Parse website with 5 seconds timeout
    Document document = Jsoup.parse(new URL(url), 5000);

    // All versions as table rows
    Elements versionRows ="table.versions > tbody > tr");

    List<MavenDependency> mvnResults = new ArrayList<>();

    // Parse a single table row
    for (Element tr : versionRows) {
      String version ="td > a.vbtn").get(0).text();
      int usages = Integer.parseInt("td:nth-last-child(3)").get(0).text());
       String type ="td:nth-last-child(2)").get(0).text();

      mvnResults.add(new MavenDependency(group, artifact, version, usages, type));

    return mvnResults;




Leave a Reply