I have played for a while writing XPath but am unable to come up with exactly what I want.
I'm trying to write XPath for link(click1 and click2 in code snippet below) based on known text(myidentity in code snippet below). Can someone take a look into and suggest possible solution?
HTML code snippet:
<div class="abc">
<a onclick="mycontroller.goto('xx','yy'); return false;" href="#">
<img src="images/controls/inheritance.gif"/>
</a>
myidentity
<span>
<a onclick="mycontroller.goto('xx','yy'); return false;" href="#">click1</a>
<a onclick="mycontroller.goto('xx','yy'); return false;" href="#">click2</a>
</span>
</div>
You don't need to use XPath here, you could use a CSS locator. These are often faster and more compatible across different browsers.
css=div:contains(myidentity) > span a:nth-child(1) //click1
css=div:contains(myidentity) > span a:nth-child(2) //click2
Note that the > is only required to workaround a bug in the CSS locator library used by Selenium.
Hard to say without seeing the rest of the HTML but the following should work:
//div[text()[contains(., "myidentity")]]/span/a
See Macro's answer - this form should be used.
//div[text()[contains(., "myidentity")]]/span/a[2]
The following only works with one section of text in the containing div.
You'll need to select based on the text containing your identity text.
Xpath for click1
//div[contains(text(),"myidentity")]/span/a[1]
Xpath for click2
//div[contains(text(),"myidentity")]/span/a[2]
Related
Can some one help me to locate an element (without using xpath) which is displayed using : <i id="ext-gen759" class="icon-tool"></i> under a <div> tag. The HTML is as follows:
<div id="ext-comp-1089" class=" MiniTbar">
<a href="javascript:void(0);" id="ext-gen760" class=" active">
<i id="ext-gen759" class="icon-tool"></i> -->> need to locate this.
</a>
</div>
I don't want to use:
By.Id --> id is dynamic
By.XPath --> not stable
I have tried the following without getting a result:
By.className("icon-tool") -- > not working
By.partialLinkText("icon-tool") --> not working
Any solution?
You can rely on the part of the id using, for example, starts-with():
//div[starts-with(#id, "ext-comp-")]/a[starts-with(#id, "ext-gen")]/i[#class="icon-tool"]
Or a CSS selector:
div[id^=ext-comp-] a.active[id^=ext-gen] i.icon-tool[id^=ext-gen]
using xpath should do this. You may need to make sure that's the only element i with same criteria on the page
//i[contains(#id,'ext-gen')]
Give a chance to the find element by css selector ?
driver.find_element_by_css_selector('i.icon-tool')
The python documentation is here http://selenium-python.readthedocs.org/en/latest/locating-elements.html
You can find it using css selector:
driver.findElement(By.cssSelector("i[class='icon-tool']"));
I have a webpage looks something like this:
<html>
...
<div id="menu">
...
<ul id="listOfItems">
<!--- repeated block start -->
<li id="item" class="itemClass">
...
<span class="spanClass"><span class="title">title</span></span>
...
</li>
<!-- repeated block end-->
<li id="item" class="itemClass">
...
<span class="spanClass"><span class="title">title something</span></span>
...
</li>
<li id="item" class="itemClass">
...
<span class="spanClass"><span class="title">title other thing</span></span>
...
</li>
</ul>
...
</div>
...
</html>
I would like to know what is the xpath of the titles ("title", "title something", "title other thing"). The point is that the order of the <li> elements are not specified. It could be different after every page loading. Is there any method how to discover a certain structure of the page with xpath? I have an notion about how to solve this issue, but before I'm going to write iterations with C# to discover the page I ask you.
Thanks in advance!
First of all, id's should be unique, so your portrayed webpage would not work well when it comes to testing.
I did however test, and got some XPath locators to work for selecting specific titles (although I recommend you fix your webpage instead of actually using this):
//li[#id='item']/span/span
//li[#id='item'][1]/span/span
//li[#id='item'][3]/span/span
If you're after all three titles, you could try Dimitre Novatchev's suggestion:
//span[#class='title']
This should get all titles on the page.
I would like to say one thing however, if you're getting into Selenium, I recommend you download the Selenium IDE extension for Firefox. It's a great tool for beginners. It helps you both to make your Selenium tests by recording your clicks on a website, and it also helps you auto-generate and test your XPath locators and other locators.
And again: I urge you to not make a website with duplicate id elements :-)
Does Selenium support XPath expressions like:
//span[#class='title']
If yes, than use the above XPath expression. It selects every span element in the XML document, whose class attribute has string value of "title".
I recommend to use a tool like the XPath Visualizer to play with different XPath expressions and see the selected nodes highlighted in the source XML document.
When I do this...
<li>
<a href="#">
<img src="#" width="#" height="#" alt="#" />
<ol class="#">
<li>#</li>
<li>#</li>
<li>#</li>
</ol>
</a>
</li>
It renders it in Firefox like this...
<li>
<a href="#">
<img src="#" width="#" height="#" alt="#" />
</a>
<ol class="#">
<li>a href="#">#</a></li>
<li>a href="#">#</a></li>
<li>a href="#">#</a></li>
</ol>
</li>
Seems to render properly in Webkit. Any ideas?
Although HTML5 now permits <a> elements to contain block-level elements (e.g. ol), Firefox's parser traditionally didn't accept that, instead converting them into a sequence of separate <a> inside the block level elements such that they only surrounded inline level elements, which is what you see.
Because Firefox was the only one of the major browser makes that did that, the Mozilla people accepted the HTML5 change, agreed to alter their parser to permit the <a> element to wrap block content. (It's just one of many parser changes for HTML5, although it seems to be possibly the most prominant one)
That change has happened in Firefox 4, so you won't see the problem there, but Firefox 3.x still uses the old behaviour.
Workarounds, include using a <div> with an onclick attribute instead of the <a>, and using JavaScript to wrap the block in an <a> element, but there's no non-JS solution. Given that (a) the page should still be usable as is, and (b) that Firefox 3.x should die out in the not too distant future, one reasonable option is just to accept the quirky Firefox 3 behaviour for now.
The <a> tag has a default style of display:inline; which makes it unsuitable for containing block level elements.
However, you can get around the problem by changing the display property of the <a> element to either block or inline-block, depending on how you want it to display.
(I note that you're using HTML5, so you'll be fine. Be aware that in xhtml, it is simply not allowed to enclose block-level elements inside an <a> tag. This won't affect you in this case, but it's worth knowing in case you ever have to work with code with an xhtml doctype)
After switching from firefox testing to internet explorer testing, some elements couldn't be found by selenium anymore.
i tracked down one locator:
xpath=(//a[#class='someclass'])[2]
While it works as it should under firefox, it could not find this element in ie.
What alternatives do i have now? JS DOM? CSS Selector? How would this locator look like?
Update:
I will provide an example to make my point:
<ul>
<li>
<a class='someClass' href="http://www.google.com">BARF</a>
</li>
<li>
<a class='someClass' href="http://www.google.de">BARF2</a>
</li>
</ul>
<div>
<a class='someClass' href="http://www.google.ch">BARF3</a>
</div>
The following xpath won't work:
//a[#class='someclass'][2]
In my understanding this should be the same as:
//a[#class='someclass' and position()=2]
and i don't have any links that are the second child of any node. All i want is, to address one link from the set of links of class 'someClass'.
Without knowing the rest of your HTML source it's difficult to give you alternatives that are guaranteed to work. Hopefully the following suggestions will help point you in the right direction:
//a[#class='someClass'][2]This is like your example, but the parantheses are not needed.
//a[contains(#class, 'someClass')][2] This will work even if the link has other classes.
css=a.someClass:nth-child(2) This will only work if the link is the 2nd child element of it's parent.
Update
Based on your update, try the following: //body/descendant::a[#class='someClass'][2]
I'm trying to automate testing of the code... well, written without testing in mind (no IDs on many elements, and a lot of elements with the same class names). I would appreciate any help (questions are below the code):
<div id="author-taxonomies" class="menu-opened menu-hover-opened-inactive" onmouseover="styleMenuElement(this)" onmouseout="styleMenuElement(this)" onclick="toggleSFGroup(this)">Author</div>
<div id="author-taxonomies-div" class="opened">
<div id="top-level-menu" class="opened">
<div id="top-level-menu-item-1" class="as-master">
<div class="filter-label"> Name</div>
</div>
<div id="top-level-menu-item-1" class="as-slave"
style="top: 525px; left: 34px; z-index: 100; display: none;"> </div>
<div id="top-level-menu-item-2" class="as-master">
<div class="filter-label">Title</div>
</div>
<div id="top-level-menu-item-2" class="as-slave">
<div id="top-level-menu-item-2" class="as-slave-title as-slave-title-subgroup"
>Title</div>
<div id="top-level-menu-item-2" class="as-slave-body"> </div>
<div class="as-slave-buffer"> </div>
</div>
<div id="top-level-menu-item-3" class="as-master">
<div class="filter-label">Location</div>
</div>
<div id="top-level-menu-item-3" class="as-slave"> </div>
</div>
</div>
The question is: how to refer particular labels of this menu and the properties with xPath expressions? For example, if I want to:
verify the "Location" label is there
check if "Title" with class "as-slave" is not visible at the moment
It would be something similar to:
//div[#id="top-level-menu-item-3"]/div[#class="filter-label"]
//div[#id="top-level-menu1"] --- and check in code for display: none ... assuming it is selenium rc you are using
Update: also be sure to install the following firefox addin, it is Really useful when trying different xpath expressions on a site https://addons.mozilla.org/en-US/firefox/addon/1095
As a side note: try to avoid using xpath locators in Selenium, if possible. If you have a long xpath expression, it can be up to 20 times slower for Selenium to find the element compared to identifying it using its unique ID. Of course, sometimes there is no alternative to using xpath. However, when you do use it, keep '//' expressions to minimum - this is a real performance killer.
If you're just starting with Selenium, download the selenium add-on for Firefox. As you click on DOM elements, Selenium shows you the xpath to access it.
I am currently working on an open source library for generating xpath expressions through a fluent .Net API. The idea is to be able to generate xpath based selenium locators without having to know xpath.
Here's an example of how the library can be used in your case:
XPathFinder.Find.Tag("div").With.Attribute("id", "top-level-menu-item-3").And.Child("div").With.Attribute("class", "filter-label").ToXPathExpression();
This will produce the following xpath:
"//div[#id='top-level-menu-item-3']/div[#class='filter-label']"
Check it out at
http://code.google.com/p/xpathitup/
You can use firepath that can be installed over firebug(both firefox plugin). When you get a xpath, dont forget to append // before using it. Either in code or in selenium IDE. You are not appending it thats why its unusable. There are two types of xpath absolute and relative. If you use absolute then it will take care of dynamic ids. But if you use relative it will break with each run.