differences between InvalidSelectorException and NoSuchElementException and locating items

Question

I am running a test nightly via Jenkins. About 95% of the time it passes. But occasionally it is failing doing this:

driver.findElement(By.xpath("//div[contains(@id, 'stage')]/div[contains(@class, 'tabMenu')]/span[contains(@class, 'menuItem') and contains(text(),'Employer')]"))

with an org.openqa.selenium.InvalidSelectorException. Note: this almost always works.

So I debug using Eclipse and Java and Selenium. Again, I had to run the test about 30 times before getting this error.

I am using Selenium and IE. IE does not have good debugging so using Eclipse Debug Shell I write the page source to a file, rename the file to an html file and edit with Chrome. In the inspect window I do a search for

 //div[contains(@id, 'stage')]/div[contains(@class, 'tabMenu')]/span[contains(@class, 'menuItem') and contains(text(),'Employer')]

and it finds it and highlights it. When I from debug shell do a

 driver.findElement(By.xpath("//div[contains(@id, 'stage')]/div[contains(@class, 'tabMenu')]/span[contains(@class, 'menuItem') and contains(text(),'Employer')]"))

I get the

 org.openqa.selenium.InvalidSelectorException:Unable to locate an element ... because of the following error:
 [object Error] (WARNING: The server did not provide any stacktrace information)

So I thought I would just get all the divs and get their IDs:

 driver.findElement(By.xpath("//div"))

and I get

 org.openqa.selenium.InvalidSelectorException: Unable to locate an element with the xpath expression //div because of the following error:
[object Error] (WARNING: The server did not provide any stacktrace information)

so OK. I try driver.findElement(By.xpath("//*")) and still get the InvalidSelectorException

so "//*" would have to find elements. And even if it didn't, shouldn't it return something like a NoSuchElementException rathe than an InvalidSelectorException? Anyone know what is happening? I did a search to find the meaning of InvalidSelectorException and found usage but not really a definition.

So I guess I have two questions: 1. why aren't any elements (even "//*") being found, and 2. Shouldn't it be returning NoSuchElementException and not InvalidSelectorExmaple?

Answer 1

NoSuchElementException

NoSuchElementException is the exception thrown by WebDriver.findElement(By by) and WebElement.findElement(By by) methods if no matching elements are found.

• NoSuchElementException extends NotFoundException which in turn extends WebDriverException.

• It's direct known subclass is InvalidSelectorException

You can find a couple of detailed discussions on NoSuchElementException in:

• NoSuchElementException, Selenium unable to locate element
• Exception in thread “main” org.openqa.selenium.NoSuchElementException: Unable to locate element: //*[@id='login-email']

---

InvalidSelectorException

InvalidSelectorException is the exception thrown when using the selector no element can be identified and as mentioned earlier is the direct known subclass of NoSuchElementException.

You can find a couple of relevant detailed discussions in:

• Invalid selector: Compound class names not permitted error using Selenium
• Selenium CSS selector throwing OpenQA.Selenium.InvalidSelectorException - what's wrong?

---

Curious case of Internet Explorer

However in a dozen of previous discussions users in favor of css-selectors say that it is more readable and faster especially when running against internet-explorer.

You can find a detailed discussion in Why should I ever use CSS selectors as opposed to XPath for automated testing?

---

This usecase

This line of code...

driver.findElement(By.xpath("//div[contains(@id, 'stage')]/div[contains(@class, 'tabMenu')]/span[contains(@class, 'menuItem') and contains(text(),'Employer')]"))

...apperantly looks perfect.

However, as per the best practices I would suggest to avoid mixing up contains() and text(). You can use either of them as follows:

• Using contains():

  driver.findElement(By.xpath("//div[contains(@id, 'stage')]/div[contains(@class, 'tabMenu')]/span[contains(@class, 'menuItem') and contains(.,'Employer')]"))
  

• Using text():

  driver.findElement(By.xpath("//div[contains(@id, 'stage')]/div[contains(@class, 'tabMenu')]/span[contains(@class, 'menuItem') and text()='Employer']"))
  

Answer 2

The InvalidSelectorException and NoSuchElementException are completely different exceptions.

• NoSuchElementException
  This is raised when locator strategy used is correct but no element matching it can be found in the page's DOM.
• InvalidSelectorException
  This is raised when the locator string that is being used is invalid. In the Selenium java implementation there are a couple of examples where this exception should be thrown:
  • testShouldThrowInvalidSelectorExceptionWhenXPathIsSyntacticallyInvalidInDriverFindElement
  • testShouldThrowInvalidSelectorExceptionWhenXPathReturnsWrongTypeInDriverFindElement

As you can see this exception is thrown when the xpath is invalid.

Why you may get InvalidSelectorException

You mentioned running your tests on IE, however IE browser doesn't have a native xpath engine for finding elements. So to be able to find elements by xpath in IE, the driver will have to use a JavaScript xpath query engine.

The above I know to be true for IE versions up until including 9. Not sure about IE10 and IE11 but you also don't mention what IE version you're using, so this may be one of the root causes.

Your xpath is syntactically correct however it may not be the case for IE.

What can you try to fix your issue

• your already making use of classes and ids so you may as well try to change selector strategy to CSS Selector. At least for the elements that are failing
• try updating your xpaths selector to be a bit less complex (if possible)
• switch to a more modern browser: Chrome, Firefox, Edge, if possible (however I doubt that anyone is using IE out of pure choice)