Question
I want to get all the links of a HTML document. This isn't a problem, but apparently it puts all the links in an alphabetic order before storing them in an array one by one. I want to have the links in original order (not in alphabetic).
So is there any possibility to get the first found link, store it, then the second one,...? I already tried using HtmlAgilityPack and the Webbrowser-Control methods, but both order them alphabetically. The original order is important for later purposes.
I heard that it might be possible with Regex, but I've found enough answers, where they say that you shouldn't use it for HTML parsing. So how can I do it?
Here's the Webbrowser-Control code, I tried to use to get the links and store them into an array:
private void btnGet_Click(object sender, EventArgs e)
{
HtmlWindow mainFrame = webFl.Document.Window.Frames["mainFrame"];
HtmlElementCollection links = mainFrame.Document.Links;
foreach (HtmlElement link in links)
{
string linkText = link.OuterHtml;
if (linkText.Contains("puzzle"))
{
arr[i] = linkText;
i++;
}
}
}
Thank you in advance, Opak
Popular Answer
You can get the correct order by walking the DOM tree using HTML DOM API. The following code does this. Note, I use dynamic to access DOM API. That's because WebBrowser's HtmlElement.FirstChild/HtmlElement.NextSibling don't work for this purpose, as they return null for DOM text nodes.
private void btnGet_Click(object sender, EventArgs e)
{
Action<object> walkTheDom = null;
var links = new List<object>();
// element.FirstChild / NextSibling don't work as they stop at DOM text nodes
walkTheDom = (element) =>
{
dynamic domElement = element;
if (domElement.tagName == "A")
links.Add(domElement);
for (dynamic child = domElement.firstChild; child != null; child = child.nextSibling)
{
if (child.nodeType == 1) // Element node?
walkTheDom(child);
}
};
walkTheDom(this.webBrowser.Document.Body.DomElement);
string html = links.Aggregate(String.Empty, (a, b) => a + ((dynamic)b).outerHtml + Environment.NewLine);
MessageBox.Show(html);
}
[UPDATE] If you really need to get a list of HtmlElement objects for <A> tags, instead of dynamic native elements, that's still possible with a little trick using GetElementById:
private void btnGet_Click(object sender, EventArgs e)
{
// element.FirstChild / NextSibling don't work because they stop on DOM text nodes
var links = new List<HtmlElement>();
var document = this.webBrowser.Document;
dynamic domDocument = document.DomDocument;
Action<dynamic> walkTheDom = null;
walkTheDom = (domElement) =>
{
if (domElement.tagName == "A")
{
// get HtmlElement for the found <A> tag
string savedId = domElement.id;
string uniqueId = domDocument.uniqueID;
domElement.id = uniqueId;
links.Add(document.GetElementById(uniqueId));
if (savedId != null)
domElement.id = savedId;
else
domElement.removeAttribute("id");
}
for (var child = domElement.firstChild; child != null; child = child.nextSibling)
{
if (child.nodeType == 1) // is an Element node?
walkTheDom(child);
}
};
// walk the DOM for <A> tags
walkTheDom(domDocument.body);
// show the found tags
string combinedHtml = links.Aggregate(String.Empty, (html, element) => html + element.OuterHtml + Environment.NewLine);
MessageBox.Show(combinedHtml);
}