Xpath using contains after an unclosed br tag

I am testing the following markup using xpath and selenium:

<table>
  <tbody>
    <tr>
      <td>
        123
        <br>
        456
      </td>
    </tr>
  </tbody>
</table>

and while this xpath works/matches:

//table/tbody/tr/td[contains(text(),'123')]

This one does not match:

//table/tbody/tr/td[contains(text(),'456')]

I'm just using "find" in chrome developer tools to test my xpath expressions so I suppose it's not a selenium question but my eventual intent is to make a function that returns true if a value is found in the table, false if not found.

My guess is that the unclosed <br> is messing things up but I dont have the ability to change the markup so I need to play the cards that I was dealt. Does anyone have any ideas of how to find out if a value exists in the table with similar markup?

I expect that I might have to get all values in every cell (which hopefully returns the stuff before and after the <br>) and then loop through them in my code.

The problem is not that the br tag is "unclosed". In fact the browser has parsed that empty <br> element and your XPath is operating over a data model which is indistinguishable from one that an XML parser would produce from either <br/> or <br></br>.

The result is that you have a td element node which contains 3 child nodes:

• a text node 123
• an empty <br> element
• another text node 456

The problem you're seeing is that the contains function is converting its first argument text() into a string, as if by a call to the string function, and the rule here is that the string value of a set of nodes is defined to be the string value of the member of that set which is first in document order; in your example that is the text node 123. So the second (and any subsequent) text node would be ignored. Once again, that would be true whether the intervening element were a br, or an img or a span or anything else.

One solution, if you want to check if the td contains the string 456, and you don't care about how that string is broken up by child elements, or whether it might be nested inside some other element (e.g. <span>456</span>) is to pass the entire string value of the td element as the first parameter to contains:

//table/tbody/tr/td[contains(.,'456')]

If you only wanted to know if the string 456 occurred in a text node which was a direct child of the td (and ignore the case where it was nested inside a span inside that td, for instance), you could use an XPath like this:

//table/tbody/tr/td[text()[contains(.,'456')]]

(Read this as "a td which has a text node child, which in turn contains the string 456")

Your assumption is correct. The <br> is causing the second locator to fail but there is a way around it.

text() gives you specifically the text inside the element itself. You can alternatively use . which basically "flattens" text inside all descendant elements. So, you can shorten both your locators to

//table/tbody/tr/td[contains(.,'123')]
//table/tbody/tr/td[contains(.,'456')]

and both will work.

NOTE: Your XPath might be overly specific. For example, by specifying the tbody layer you are excluding table headings. If you ONLY want to search the body of the table, then you're good.