TThier/Languages/xml/xpath
Dan Bikle


This page contains a few simple XPath demonstrations.

XPath needs something to chew on.  A simple HTML file works pretty well for this.
A simple HTML file is displayed below.

Rendered:
simple.html

Text:
simple-html.txt

XPath has an easier time chewing on HTML if we force the HTML to be "well formed".

We may use Tidy.jar to transform an HTML file into a well formed HTML file.

Here is a command line which demonstrates how to use Tidy.jar
$JAVA_HOME/bin/java -classpath Tidy.jar org/w3c/tidy/Tidy -indent -wrap 999 -asxml -utf8 simple.html > simpleWellformed.html

Here is a demo shell script which wraps the above command line:
Tidy-sh.txt

Here are some links to Tidy.jar:
Tidy.jar
http://jazilla.sourceforge.net/products/jars/Tidy.jar
http://prdownloads.sourceforge.net/httpunit/httpunit-1.5.4.zip?download (Tidy.jar is inside this .zip)

After I made simple-html.txt well formed with Tidy.jar,
I ended up with this:

Rendered:
simpleWellformed.html

Text:
simpleWellformed-html.txt

The demonstrations below help you translate XPath statements into
English.  If you already know some HTML authoring, the demonstrations
might pack more punch since they connect with some of your HTML syntax knowledge.


English to XPATH translations





English:
Give me a list of all the <a> elements.


XPath:
//xhtml:a



Notice that we need to specify the namespace via the "xhtml" namespace prefix.
Be aware that XPath is very picky about requiring us to specify a namespace.







The above XPath expression would return the following syntax to XSLT or XQuery:
allA.txt






English:
Give me a list of all the <a> elements which reside in a <td> element
which resides inside a <tr> element which resides in a <body> element
which  which resides in an <html> element.


XPath:
/xhtml:html/xhtml:body/xhtml:table/xhtml:tr/xhtml:td/xhtml:a







The above XPath expression would return the syntax below to XSLT or XQuery:


h-b-t-tr-td-a.txt







We can see that the addition of the xhtml namespace by Tidy.jar complicates
both the XPath expression and the elements returned by it.

In order to simplify this demonstration, I did some editing, by hand, to remove
the xhtml namespace from simpleWellformed.html:

Rendered:
simpleWellformedNoNamespace.html

Text:
simpleWellformedNoNamespace-html.txt

Now, we return to the translations.

More English to XPATH translations (without namespaces)


  

    English: Give me a list of all the <a> elements which reside in a <td> element which resides inside a <tr> element which resides in a <body> element which resides in an <html> element.
    XPath:
/html/body/table/tr/td/a

      

      

        The above XPath expression would return the following syntax to XSLT or XQuery: h-b-t-tr-td-a-nns.txt 
      

      

        English: Give me a list of all the "href" attributes inside of <a> tags.
        XPath:
//a/@href

      

      

        The above XPath expression would return the following syntax to XSLT or XQuery: allHref.txt 
      

      

        English: Give me a list of all <a> elements which contain an "id" attribute.
        XPath:
//a[./@id]

      

      

        
          The above XPath expression would return the following syntax to XSLT or XQuery: aWithId.txt.
          I like to think of the [ ] syntax as like a box.  I use this thought pattern: "Look inside all the boxes;
          if any of the boxes contain candy, return those boxes and ignore the rest."  Here is another way to put it:
          "/box/candy" returns candy with no box. /box[candy] returns boxes which contain candy."
        
      

      

        English: Give me a list of all <a> elements which do not contain an "id" attribute.
        XPath:
//a[not(./@id)]

        		
      

      

        The above XPath expression would return the following syntax to XSLT or XQuery: aNotWithId.txt 
      


      

        English: Give me a list of all <td> elements which contain an <a> element.
        XPath:
//td[a]

        		
      


      

        
          The above XPath expression would return the following syntax to XSLT or XQuery: tdWithA.txt
        
      

      
      

        English: Give me the text from all <td> elements.
        XPath:
//td/text()

        		
      

      

        
         The above XPath expression would return the following syntax to XSLT or XQuery: tdText.txt. 
         Some text in an element is often referred to as a "text node".
        
      


      

      English: Give me a list of  all <a> elements which contain an "href" attribute which ends with "gov".
      
      XPath:
//a[ ends-with(./@href,"gov")) ]

      		
      

      

      The above XPath expression would return the following syntax to XSLT or XQuery:
        endsWithGov.txt.
      
      


      

      English: Give me a list of  all <a> elements which do not contain an "id" attribute
          and also do not contain an "href" attribute which ends with "gov".
      
      XPath:
//a[not(./@id) and not(ends-with(./@href,"gov")) ]

      		
      

      

      The above XPath expression would return the following syntax to XSLT or XQuery: 
notEndsWith.txt.
      
      


      

      English: Give me the text nodes from all the <a> elements which do not start with "http".
      
      XPath:
//a[ not (starts-with(./text(), "http")) ]/text()

      		
      

      

      The above XPath expression would return the following syntax to XSLT or XQuery:
        notStartsWithText.txt
      
      


      

      English:Give me a list of  all <a> elements which contain a text node which is exactly
             equal to "Stanford".
      
      XPath:
//a[ ./text() = "Stanford" ]

      		
      

      

      The above XPath expression would return the following syntax to XSLT or XQuery: 
          aWithTextEquals.txt
      
      



      

      English:  Give me a list of all text nodes which reside in <a> elements which contain a text node which is exactly
             equal to "Stanford".  In other words,
             if any <a> elements contain a text node equal to "Stanford", give me those text nodes.
      
      XPath:
XPath: //a[ ./text() = "Stanford" ]/text()

      		
      

      

      The above XPath expression would return the following syntax to XSLT or XQuery: 
          textOfaWithTextEquals.txt
      
      


  


I typed up a few more translations and linked them below in a text file:
moreTranslations.txt
yetMore.txt