Navigation in XQuery

This chapter is from the book

XQuery: The XML Query Language

Learn More Buy

This chapter is from the book

This chapter is from the book 

XQuery: The XML Query Language

Learn More Buy

3.5 Navigation Examples

To illustrate the navigation concepts introduced in this chapter, let's consider a variety of different navigation tasks over the sample XML document, team.xml, from Chapter 1. For convenience, it's repeated in Listing 3.6.

This document contains employee information from a fictitious organization. The data consists primarily of Employee elements, in which parent/child relationships in the XML correspond to manager/employee relationships in the organization. Just to spice things up a bit, the document also contains a few comments and processing instructions.

Listing 3.6. The team.xml document

<?xml version='1.0'?>
<Team name="Project 42" xmlns:a="urn:annotations">
  <Employee id="E6" years="4.3">
    <Name>Chaz Hoover</Name>
    <Title>Architect</Title>
    <Expertise>Puzzles</Expertise>
    <Expertise>Games</Expertise>
    <Employee id="E2" years="6.1" a:assigned-to="Jade Studios">
      <Name>Carl Yates</Name>
      <Title>Dev Lead</Title>
      <Expertise>Video Games</Expertise>
      <Employee id="E4" years="1.2" a:assigned-to="PVR">
        <Name>Panda Serai</Name>
        <Title>Developer</Title>
        <Expertise>Hardware</Expertise>
        <Expertise>Entertainment</Expertise>
      </Employee>
      <Employee id="E5" years="0.6">
        <?Follow-up?>
        <Name>Jason Abedora</Name>
        <Title>Developer</Title>
        <Expertise>Puzzles</Expertise>
      </Employee>
    </Employee>
    <Employee id="E1" years="8.2">
      <!-- new hire 13 May -->
      <Name>Kandy Konrad</Name>
      <Title>QA Lead</Title>
      <Expertise>Movies</Expertise>
      <Expertise>Sports</Expertise>
      <Employee id="E0" years="8.5" a:status="on leave">
        <Name>Wanda Wilson</Name>
        <Title>QA Engineer</Title>
        <Expertise>Home Theater</Expertise>
        <Expertise>Board Games</Expertise>
        <Expertise>Puzzles</Expertise>
      </Employee>
    </Employee>
    <Employee id="E3" years="2.8">
      <Name>Jim Barry</Name>
      <Title>QA Engineer</Title>
      <Expertise>Video Games</Expertise>
    </Employee>
  </Employee>
</Team>

Each Employee has an id attribute that we will assume has been typed as xs:ID with a DTD or XML Schema so that it can be looked up by the id() lookup function. And finally, the team.xml document contains some “annotations” in another namespace ("urn:annotations"). These attributes describe additional information about the employees and are used here to demonstrate navigation using qualified names and the other wildcard node tests.

For the first example, the team.xml document is loaded using the doc() function. For the remaining examples, we will assume that this document is already the input sequence, so that all paths are resolved relative to it without loading it explicitly.

As our first task, consider finding the names of all employees. Because Employee elements occur at many different levels in the XML, use the descendant navigation shortcut // to match every Employee element descendant of the root document node. Finally, select their child elements named Name. The result is a list of the names of all employees in the document (returned in document order), as shown in Listing 3.7.

Listing 3.7. Find the names of all employees

doc("team.xml")//Employee/Name
=>
<Name>Chaz Hoover</Name>
<Name>Carl Yates</Name>
<Name>Panda Serai</Name>
<Name>Jason Abedora</Name>
<Name>Kandy Konrad</Name>
<Name>Wanda Wilson</Name>
<Name>Jim Barry</Name>

Suppose instead we want to select only some of the employees, subject to some condition as in Listing 3.8.

Listing 3.8. Name all employees who have been at the company less than two years

//Employee[@years < 2]/Name
=>
<Name>Panda Serai</Name>
<Name>Jason Abedora</Name>

This path is the same as the previous one, except that a predicate has been added to the Employee step. We want to filter the Employee elements so that we select only those whose years attribute has a value less than 2. So we use the attribute axis @ and the less-than comparison operator < to compare the years attribute against 2. Then from these filtered employees, their names are selected as in the previous example.

We can also search for attributes in another namespace. For example, we could search for all employees currently assigned, as shown in Listing 3.9.

Listing 3.9. Find all employees currently assigned

declare namespace ann = "urn:annotations";
//Employee[@ann:assigned]/Name
=>
<Name>Carl Yates</Name>
<Name>Panda Serai</Name>

In this query, we have used the query prolog to declare a namespace prefix, and then used this prefix in the attribute name test @ann:assigned to match attributes with the local name equal to assigned and namespace equal to urn:annotations. Note that the prefix used in the query can be (and in this case is) different from the one used in the original document.

By putting the attribute test in the predicate with no comparison, we test for its existence. The predicate is converted using the Effective Boolean Value rule, which tests whether the sequence is non-empty. Consequently, this XPath finds all employees with an assignment, regardless of what that assignment actually is. Similarly, we could find all employees who lack an assignment by applying the not() function, as shown in Listing 3.10.

Listing 3.10. Find all unassigned employees

declare namespace ann = "urn:annotations";
//Employee[not(@ann:assigned)]/Name
=>
<Name>Chaz Hoover</Name>
<Name>Jason Abedora</Name>
<Name>Kandy Konrad</Name>
<Name>Wanda Wilson</Name>
<Name>Jim Barry</Name>

The query to find all employees with an expertise in puzzles is superficially similar to the previous query. The previous query needed to compare attribute values; this query (see Listing 3.11) needs to compare child element values.

Listing 3.11. Find all employees skilled in puzzles

//Employee[Expertise = "Puzzles"]/Name
=>
<Name>Chaz Hoover</Name>
<Name>Jason Abedora</Name>
<Name>Wanda Wilson</Name>

This case is made somewhat more difficult by the fact that employees may have more than one expertise. Consequently, we must test whether there exists any child expertise element with the desired value. Fortunately, the general comparison operators like < and = are defined so that they already do this existence test implicitly (see Chapter 5). Thus, the predicate Expertise = "Puzzles" tests whether there exists a child element named Expertise whose string value is "Puzzles".

Navigation can also be used to compute other values. For example, Listing 3.12 counts the number of employees in one division.

Listing 3.12. Count the number of people in Chaz Hoover's organization

count(//Employee[Name="Chaz Hoover"]/descendant-or-self::Employee)
=>
7

The count() function computes the number of items in a sequence (see Chapter 5). First we must locate the employee named Chaz Hoover, which can be done using a query like the ones used previously. But then we must count all the employees contained in the sub-tree rooted at Chaz Hoover—in other words, all the descendant Employee elements. By using the descendant-or-self axis, we have included Chaz Hoover himself in this count. We could exclude him by instead using the descendant axis as in the path count(//Employee[Name="Chaz Hoover"]/descendant::Employee).

Instead of counting the entire organization, we could instead count only those employee elements directly under Chaz Hoover, as shown in Listing 3.13.

Listing 3.13. Count the number of Chaz Hoover's direct reports

count(//Employee[Name="Chaz Hoover"]/Employee)
=>
3

Instead of performing a descendant query with //, we have used ordinary child navigation / to select only those employees who report directly to Chaz Hoover. This task could also be accomplished in another way, using the parent axis, as shown in Listing 3.14.

Listing 3.14. Use the parent axis to count Chaz Hoover's employees

count(//Employee[../Name="Chaz Hoover"])
=>
3

Here we first find every Employee in the document. Then, we check to see if the parent element has the name Chaz Hoover. We use the .. abbreviation to navigate to the parent, and then compare its child Name element. This query is usually much slower than the previous one, although some implementations can optimize it so that both perform identically.

We can also find other kinds of elements in the tree. For example, we could extract all comment and processing-instruction nodes by using node kind tests, as illustrated in Listing 3.15.

Listing 3.15. Find all comments and processing instructions

//comment() | //processing-instruction()
=>
<?Follow up?>
<!-- new hire 13 May -->

In this query, we used the union operator | to combine the results of both paths. We could have also written //(comment() | processing-instruction()) to achieve the same effect. (See Chapter 5 for more information about the union operator.)

Notice that the XML declaration <?xml version='1.0'?> at the top of the document did not match. Although it looks like a processing instruction, XML doesn't treat it as part of the data model, so XQuery doesn't either. Finally, Listing 3.16 demonstrates looking up elements by their IDs.

Listing 3.16. Find all employees with the same job function as employee E0

//Employee[Title = id("E0")/Title]/Name
=>
<Name>Wanda Wilson</Name>
<Name>Jim Barry</Name>

Because we wish to find all employee names satisfying some condition, we know that the path will consist of //Employee/Name and use a predicate to limit which employees are matched. This predicate should select all employees with the same title as that of employee E0. Employee E0 can be found using the id() navigation function: id("E0"). Then all that remains is to compare the current employee's title against that of E0.

Notice that instead of using id(), we could use an absolute path inside the predicate to search from the root of the document to find the employee with id E0: //Employee[Title = //Employee[@id="E0"]/Title]/Name. This path has the advantage that it doesn't require a DTD or schema to type the id attribute. However, it is more complex to write and usually will perform worse than the id lookup (which most implementations optimize into an index or table lookup). This path essentially performs a join of the document with itself. Joins like this are often expressed using FLWOR expressions (described in Chapter 6).