How do I select an empty element in XPath?

Question

We select elements with Django 1.4 tests and Selenium like this:

self.assertEqual(1, len(self.selenium.find_elements_by_xpath("//a[@href='#'][@class='button save-as'][@title='Save As...'][text()='Save As']")))

(The class inherits from LiveServerTestCase).

The problem is that sometimes there are elements without text, and if we select with [text()=''] it fails (the len is 0). How can I select elements without text?

Update: Because [text()=''] didn't work, I had to assert two lines to assert no text:

self.assertEqual(1, len(self.selenium.find_elements_by_xpath("//a[@href='#'][@class='button properties'][@title='Properties']")))
self.assertEqual("", self.selenium.find_element_by_xpath("//a[@href='#'][@class='button properties'][@title='Properties']").text)

Now I can assert the same with one line:

self.assertEqual(1, len(self.selenium.find_elements_by_xpath("//a[@href='#'][@class='button properties'][@title='Properties'][not(text())]")))

Solution

You could use XPath's not() function.

//a[@href='#'][@class='button save-as'][@title='Save As...'][not(text())]

Comments

[Uri]

It works, just one fix - it's [not(text())] and not [text()='Save As' or not(text())] - there is only one option of no text.

[Matt Deacalion]

Ah, so you know before hand that the elements should be empty? I read your question as you weren't sure what the element contained. Edited.

[Uri]

Yes, some elements contain text and some are empty. They are different elements.

[Uri]

By the way, do you know why [not(text())] works and [text()=''] doesn't?

[Matt Deacalion]

I'm not sure, I pulled this out of some of my old code. I think it has something to do with an empty node (no children) and a text node containing an empty string (children) being treated differently in XPath.