Home/ Questions/Q 8664341
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-06-12T17:13:06+00:00

I need to extract the text from the nodes in an html file and

  • 0

I need to extract the text from the nodes in an html file and I’m trying to use XPath and Javascript.

The required condition is that the text must contain an specific word.

Let’s take by example the next html file:

<html>
    <body>
        <p>
            Hi, try to extract the word username here and here <b>username</b>
        </p>
    </body>
</html>

And try to get the text from text nodes containing the word ‘username’ with this expression:

var search = document.evaluate('//*[contains(child::text(), \"username\")]/child::text()', document, null, XPathResult.ANY_TYPE, null);

Iterating through search I’ve found the desired result but unwanted objects too:

["Hi, try to extract the word username here and here", Text, "username"]

where Text is an Object whose textContent is only the carriage return symbol (I’m using Google Chrome console). Where does this object come from?

Can anyone, please, give a more precise XPath expression that excludes those Objects or should I exclude them in my code?

The ideal search should be:

["Hi, try to extract the word username here and here", "username"]

Thanks everybody!

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    Looks like you want

    var search = document.evaluate('//text()[contains(., \"username\")]',
   document, null, XPathResult.ANY_TYPE, null);

    (I’m not sure why you’re escaping your double quotes inside single quotes, but that’s a separate issue.)

    Your existing code,

    var search = document.evaluate('//*[contains(child::text(), \"username\")]/
         child::text()', document, null, XPathResult.ANY_TYPE, null);

    says,

    • (1) Find all elements that have text node children that contain “username”.
    • (2) Now return all the text node children of those elements (regardless of what they contain).

    The (1) part of the expression will return element nodes <p> and <b>.

    For the (2) step, <b> has only one text node child, but <p> has two: the one before the <b> (which contains “username”) and the one after the <b> (which contains only whitespace).

    Solution: Forget about elements — they are an irrelevant distraction. Just select the desired text nodes directly.

    • 0