Let's spend a few minutes on a few basic terms commonly used. You have seen that I use the term XPath expression and XPath query interchangeably. An XPath expression is also called a "Location Path," because you are selecting one or more locations in your XML document and, to be able to do so, you specify the path to it. The expression "//Employees/Employee" is a path to certain locations in your XML document you want to select. It is the path to these locations. The "Location Path" consists of "Location Steps."
Your path has two steps: First, go to the Employees node and then, from there, go to all the Employee nodes. Each "Location Step" consists of an axis, a node-test, and a predicate or filter. The axis defines the relationship between your current location and the next location you specify. When no location is specified (see table below), child is assumed. The node-test is the node you want to select. You provide the name of the node or the wildcard operator for any node. The predicate or filter is any filter you apply by using the filter operator. The expression //child::Employees/child::Employee" is the same as "//Employees/Employee". Let's look closer at some of the most common axes:
Axes
Description
ancestor::
This returns all ancestors of the selected node. The expression "/Employees/Employee/JobTitle/ancestor::*" selects the JobTitle of each Employee node and then returns all its ancestors up to the root, so it returns the Employee node with the ID one, the Employee node with the ID two, and the Employees node itself. This makes it easy to find all ancestors of your nodes. But, you also can look for a specific ancestor; for example, with the expression "//JobTitle[ancestor::Employee/@ID=1]", you select all JobTitle nodes that have an ancestor node Employee with an ID attribute of the value one.
child::
This axis is the default if none is specified and has the same result as the child operator. For example, the expression "/child::Employees/child::Employee" selects all the Employee nodes in the Employees node.
self::
This has the same result as the current context operator. The expression "//*/self::JobTitle" first selects all nodes in the XML document and then refers to the current context and from there selects only the nodes of the type JobTitle. The expression "//*/./JobTitle" returns the same result using the current context operator.
ancestor-or-self::
Same as the ancestor axis, but it includes the current context. The expression "//JobTitle/ancestor-or-self::*" lists both Employee nodes, both JobTitle nodes, and the Employees node. But, on the other hand, the expression "//JobTitle/ancestor::*" does not return the two JobTitle nodes.
attribute::
The attribute axis returns the same as the attribute operator; it walks the attributes instead of the elements. The expression "//attribute::*" as well as "//@*" return all the attributes in your XML document.
descendant::
The descendant operator returns the same as the recursive descendant operator. It returns all descendant nodes. The expression "//Employee/descendant::*" as well as "//Employee//*" return all descendant nodes under the two Employee nodes.
descendant-or-self::
Same as the descendant axis but it includes the current context. So, the expression "//Employee/descendant-or-self::*" lists both Employee nodes, including all its descendant nodes. But, on the other hand, the expression "//Employee/descendant::*" does not return the two Employee nodes.
following::
Selects all the nodes following the current context node, which includes all siblings (nodes at the same hierarchy level) and its children but not their descendants. The expression "//Employee[@ID=1]/following::*" selects the Employee node with the attribute ID value of one and then returns all its following Employee nodes, in our case only the one with the attribute ID value of two, and its children, but not the descendants of the children.
following-sibling::
Returns the same as the following axis but includes only siblings and no children and not their descendants. So, the expression "//Employee[@ID=1]/following-sibling::*" returns only the Employee node with the attribute ID value of two.
preceding::
Selects all the nodes preceding the current context, which includes all siblings (nodes at the same hierarchy level), and its children but not its descendants. The expression "//Employee[@ID=2]/preceding::*" selects the Employee node with the attribute ID value of two and then returns all its preceding Employee nodes—in our case, only the one with the attribute ID value of one, and its children but not the descendants of the children.
preceding-sibling::
Returns the same as the preceding axis but includes only siblings and no children and not its descendants. So, the expression "//Employee[@ID=2]/preceding-sibling::*" returns only the Employee node with the attribute ID value of one.
The axes give you more control over which nodes you want to select. It makes it easy to find ancestors, descendants, following, and preceding nodes. Unfortunately, you cannot not run XPath queries against an XML document from within the Visual Studio .NET 2003 IDE. The overall support of XPath is rather weak in Visual Studio.
You can, however, use commercial tools such as XML Spy from Altova (http://www.altova.com), Stylus Studio from Sonic Software (http://www.stylusstudio.com), or as I use a simple freeware tool called "Visual XPath" (http://weblogs.asp.net/nleghari/articles/27951.aspx). It is very useful to use such tools so you can experiment with XPath and easily create the right XPath query you need plus visually verify that you got the expected result. Here are a few more examples to study:
Sample 1
/descendant::JobTitle[ancestor::Employee/@ID=2]
Selects all descendant JobTitle nodes that have as an ancestor an Employee node with the attribute ID value of two. The result will be as follows:
<JobTitle>
Sr. Developer
</JobTitle>
Sample 2
//Employee[WebAddress and LastName='Salchner']/attribute::ID
Selects all descendant Employee nodes that have a WebAddress child node and a LastName child node with the value Salchner. From the result, it selects the ID attribute, which returns the values of the selected attributes. The result returns 1, the value of the selected ID attribute.
Sample 3
//JobTitle[ancestor::*/@ID=1]
Selects all descendant JobTitle nodes that have any ancestor node with an attribute ID value of one. The result is as follows:
<JobTitle>
Sr. Enterprise Architect
</JobTitle>
Understanding the base operators, functions, and axes allows you to build very powerful XPath queries and easily select the XML elements, attributes, or nodes you require. In your filter, you can use a number of boolean and comparison expressions as well as string functions. Here is a list of the most common ones:
Expression or Function
Description
and (logical and)
Allows you to have two or more logical 'and' conditions in your filter. For example, the expression "//Employee[FirstName='Klaus' and LastName='Salchner']" checks for both the FirstName and LastName.
or (logical or)
Allows you to have two or more logical 'or' conditions in your filter. The expression "//Employee[FirstName='Klaus' or FirstName='Peter']" checks for nodes that have as FirstName either Klaus or Peter.
= (equal operator)
For example, "//FirstName[.='Klaus']".
!= (unequal operator)
For example, "//FirstName[.!='Klaus']".
not() (not operator)
For example, "//Employee[not(FirstName='Klaus')]".
"<=" or "<=" (less then equal operator)
For example, "//Employee[@ID<=1]".
"<" or "<" (less then operator)
For example, "//Employee[@ID<2]".
">=" or ">=" (greater equal operator)
For example, "//Employee[@ID>=2]".
">" or ">" (greater operator)
For example, "//Employee[@ID>2]".
position()
Returns the position or index relative to all the selected nodes. The expression "//Employee[position()=1]" returns the first Employee node.
last()
Returns the position or index of the last of all selected nodes. The expression "//Employee[position()=last()]" returns the last Employee node.
concat()
Concatenates two strings together. You can specify two string literals or actual node names. The following expression, //Employee[concat(FirstName,LastName)=concat('Klaus','Salchner')], concatenates the FirstName and LastName together and compares them to the concatenated string literals 'Klaus' and 'Salchner'. All matching nodes are returned.
contains()
Checks whether the first string contains the second string. You can specify node names or string literals. The following expression, "//Employee[contains(FirstName,'Kl')]", returns all nodes containing the string 'Kl' in the FirstName node.
starts-with()
Checks whether the first string starts with the second string. You can specify node names or string literals. The following expression, "//Employee[starts-with(FirstName,'Kl')]", returns all nodes starting with the string 'Kl' in the FirstName node.
substring-after()
Checks whether the first string contains the second string and returns the remaining string after the first occurrence. The expression "//Employee[substring-after(FirstName,'Kl')='aus']" first looks whether the string 'Kl is included in the FirstName node and returns the remaining string after the first occurrence. The FirstName node containing 'Klaus' contains the string 'Kl' and the remaining string returned is 'aus'. It selects all nodes that match this criteria.
substring-before()
Checks whether the first string contains the second string and returns the preceding string before the first occurrence. The expression "//Employee[substring-before(FirstName,'aus')='Kl']" first looks whether the string 'aus is included in the FirstName node and returns the preceding string before the first occurrence. The FirstName node containing 'Klaus' contains the string 'aus' and the preceding string returned is 'Kl'. It selects all nodes which match this criteria.
substring()
Returns a sub-string starting at the position specified and with the number of characters specified. This allows you to cut out a sub-string from the specified string. You can specify a node or string literal. The expression "//Employee[substring(FirstName,2,3)='lau']" takes the node FirstName and returns a sub-string starting at position two for three characters in length. The FirstName node containing 'Klaus' returns the sub-string 'lau'. It selects all nodes that match this criteria.
string-length()
Returns the length of a string. You can give a node name or string literal. The expression "//Employee[string-length(LastName)>=8]" returns all nodes that have a LastName node of greater or equal of eight characters.
The table does not contain a complete list but explains the most common expression and functions. For a complete list, please refer to the XPath standard.
Summary
XPath queries are a very efficient and powerful way to query and navigate your XML documents. The operators, axes, and functions provided by XPath allow you to create very simple and complex queries. This will always be more efficient then navigating the XML documents on your own. XPath queries are also very important in writing XSLT transformations so you can select the right data to output and format. For a complete reference, refer to the W3C Web site or your MSDN library.
The next article in this series will explain the fundamentals of XSD schemas. If you have comments to this article, please contact me at klaus_salchner@hotmail.com or post at the bottom of this article. I want to hear if you learned something new. Contact me if you have questions about this topic or article.
About the Author
Klaus Salchner has worked for 14 years in the industry, nine years in Europe and another five years in North America. As a Senior Enterprise Architect with solid experience in enterprise software development, Klaus spends considerable time on performance, scalability, availability, maintainability, globalization/localization, and security. The projects he has been involved in are used by more than a million users in 50 countries on three continents.
Klaus calls Vancouver, British Columbia his home at the moment. His next big goal is running in the New York marathon in 2005. Klaus is interested in guest speaking opportunities or as an author for .NET magazines or Web sites. He can be contacted at klaus_salchner@hotmail.com or http://www.enterprise-minds.com.
Add www.developer.com to your favorites Add www.developer.com to your browser search box IE 7 | Firefox 2.0 | Firefox 1.5.xReceive news via our XML/RSS feed
IE 7
Firefox 2.0
