XPath query to select all href attributes of <a> tag, which 'class' attribute equals specified string

Question

I don't know why following query doesn't work:

//a/@href[@class='specified_string']

Solution 1

Try it the other way round:

//a[@class='specified_string']/@href

After all, class is an attribute of the <a> element, not an attribute of the href attribute.

Solution 2

An attribute cannot have attributes. Only elements can have attributes.

The original XPath expression:

//a/@href[@class='specified_string'] 

selects any href attribute of any a element, such that the href attribute has an attribute class whose value is 'specified_string'.

What you want is:

//a[@class='specified_string']/@href 

that is: the href attribute of any a element that has class atribute with value 'specified_string'.

Solution 3

You basically say that you are looking for an attribute named href, whose attribute (this is the error) class should be equal to specified_string.

But you need to find the attribute href of an element a, whose attribute class is specified_string.

(ndim's answer overlapped mine)

Comments

[kans]

As the class attribute may contain multiple class names separated by spaces, you probably actually want: //a[contains(concat(' ',normalize-space(@class),' '), 'some_class_name')]/@href

[kmkaplan]

@singpolyma: Good point. Just a nitpick: it is ' some_class_name ' (with spaces around), not 'some_class_name'.

[l245c4l]

Yes I think it works, but the problem is that query returns attributes, and I want to retrieve attributes' values. I'm writing it in Java and it returned me NodeList of length 2. When I try to print it out, it just prints two xml headers: <?xml version="1.0" encoding="UTF-8"?><?xml version="1.0" encoding="UTF-8"?> two and there is only two links that query will match, so it seems working. But I want to get values of this hrefs. How to do that?

[ndim]

That NodeList should contain the list of Nodes representing the href attributes. So just running getNodeValue() on those nodes should give you the href attributes' values.

[Skywalker326]

This is not working in my case, although I think it's totally right. My Chrome can use //a[@name= "topic"] to extract the right a elements. But //a[@name= "topic"]/@href returns a list of [, , , , , ] as if there isn't a href attribute. This is the an "a element" as example: <a href="/topic/19778287" name="topic">XXX</a>

[Skywalker326]

Just to let you guys know that that Xpath actually worked. But Chrome console somehow displays an empty-like list. When I assign that list to a variable and print that variable, everything I want is there. I am leaving my original comment there so other people with same issue can relate.