See the fruit of my labor, and if you want to save yourself a frustrating afternoon of getting quotes and parens to correctly match, I recommend you not modify it much. XPATH 1.0 can’t used escaped quotes like Java can, it relies on use of concatenate function to get things working in that case. So, we’re limited to using XPATH 1.0’s translate function which has its own issues with special characters, namely quotes. Typically, XPATH 2.0 has a method that takes care of this, but our framework is only capable of using XPATH 1.0 as of now. Multiple spaces that are really half spaces, capital Ys that are actually similar looking Greek counterparts, etc. After ripping out a few hairs, we discovered that the app’s encoding is in UTF-8, and is full of similar-but-not-really-the-same characters. The text exactly matches, but the XPATH won’t execute correctly. While we start simple, as we progress further through the application, these paths keep breaking for seemingly no reason. This is especially useful for writing tests for this application, where a lot of our XPATHs need to be able to select unique options on the page to interact with them. Though irritating when this text changes, we have created unique selectors out of nothing.

Then the header with text '

' appears

These elements lack any ids or other defining attributes that would make them easy to spot, and not wanting to build fragile, dynasty based xpath selectors, we’ve settled on using the text itself as our unique identifiers.

Work at a current client has led to – for a various amount of restrictions with reasons that I won’t get into – the need for using dynamically generated XPATHs in order to locate elements on the page.

If you’re here because you now have to fix the method that had this link commented in it, I’m sorry.

How to write xpath for text