xpath to get Node containing text

Question

I tried to search for nodes containing text 'Yahoo' under '/doc/story/content', it returns 'content' node, but I need exact text node that contains 'Yahoo' or it's parent

<doc>
    <story>
        <content id="201009281450332423">
            <ul>MSW NYNES NYPG1 DILMA</ul>
            <p> <k> Yahoo, made </k> it nice </p>
            <p>
               <author>-v-</author>
            </p>
        </content>
    </story>
</doc>

Xpath: "/doc/story/content[contains(., 'Yahoo')]"

Solution 1

Since you need all textNodes only which contain the text Yahoo, use the following XPath.

//text()[contains(., 'Yahoo')]

This should return you all the textNodes only which contains Yahoo (case-sensitive) in it.

Solution 2

Your XML is malformed. </content></doc></story> should be </content></story></doc>.

Apart from that, the XPath you would want is

/doc/story/content//*[contains(., 'Yahoo')]

(select any descendant of <content> which contains the text "Yahoo" -- this will select the <p>)

Comments

[Vjy]

This works great if it's one level down, How to make it work for multi-nested tags?

[Waihon Yew]

@Vjy: I 'm not sure what you mean. Can you give an example?

[Vjy]

Updated the above xml with additional tag <K>, it should select K instead of P tag. this is just example, the text node can be n level deep.

[Emiliano Poggi]

@Vjy: this does exactly what you asked for.

[Jason S]

text() is a node test not a string. contains() expects strings. See stackoverflow.com/a/9493870/695671 Your solution may appear to work, but I have a case with text nodes within text nodes in which case it fails.

[Waihon Yew]

@JasonS: That situation did not cross my mind (how did you manage to do it? programmatically?). I have corrected the answer accordingly. Thank you for pointing that out, I feel I learned something new.

[Jason S]

@Jon I did it as in your updated answer. I am getting content from text nodes in odt files using PHP SimpleXMLElement. The odt often has paragraphs with tabs and spaces represented like <text:p Hello<text:s/><text:tab/>Jon</text:p>, in which case searching using contains(text(),"Jon") will fail, but contains(.,"Jon") will work.

[Nakilon]

What it the difference between this answer and @Jon's?

[Stefan Steiger]

Case insensitive: //text()[contains(translate(., 'ABCDEFGHIJKLMNOPQRSTUVWXYZÄÖÜÉÈÊÀÁÂÒÓÔÙÚÛÇÅÏÕÑŒ', 'abcdefghijklmnopqrstuvwxyzäöüéèêàáâòóôùúûçåïõñœ'),'yahoo')]