Tricks with XPath

Example – 1:

<div id="first-div">

<div id="text-1" name="A">Text</div>

<div id="text-2" name="B">Random text</div>

<div id="text-3">Random text</div>

</div>

If you need element and it has id, standard text or name, it’s easy to find it by id or name:

.//div[@id=’text-1’]

.//*[name()=’A’]

 

By text:

.//*[contains(text(),’which will never change’)]

OR

.//*[contains(.,’which will never change’)]

OR

.//*[starts-with(.,’Text’)]

OR

.//*[ends-with(.,’change’)]

 

If you know full text:

.//*[text()=’Text for testing’]

OR

.//*[.=’Text for testing’]

 

Perhaps you want to find all divs or all elements which haven’t random text:

.//div[not(.=’Random text’)]

 

But now you need all elements which haven’t id and have random text, so:

.//div[.=’Random text’ and not(@id)]

 

 

Example – 2:

<div class="ui-dialog ui-widget" style="position: absolute; height: auto; width: 800px; top: 191px; left: 557px; display: block;" tabindex="-1">
 <div class="ui-dialog-titlebar">
 <span id="random1223">Window</span>
 <button>
 <span class="ui-button" />
 <span class="ui-button-text">New</span>
 </button>
 </div>
 ....
 <div class="uidialog-buttonpane">
 <div class="ui-dialog-buttonset">
 <button>
 <span class="ui-button-text">New<span>
 </button>
 </div>
 </div>
</div>

We can find main div by class attribute, but in case class changes, we need to rewrite our xpath.

BAD:      .//span[@class=’ui-button-text’]

BAD:      .//span[2][@class=’ui-button-text’]

You can see that span on the footer of the Window doesn’t have siblings, so we have rule which can be implemented: .//div[div[span=’Window’]] and find span with text() equals New.

.//div[div[span=’Window’]]//span[.=’New’]