pugixml 1.4 manual | Overview | Installation | Document: Object model · Loading · Accessing · Modifying · Saving | XPath | API Reference | Table of Contents


XPath types
Selecting nodes via XPath expression
Using query objects
Using variables
Error handling
Conformance to W3C specification

If the task at hand is to select a subset of document nodes that match some criteria, it is possible to code a function using the existing traversal functionality for any practical criteria. However, often either a data-driven approach is desirable, in case the criteria are not predefined and come from a file, or it is inconvenient to use traversal interfaces and a higher-level DSL is required. There is a standard language for XML processing, XPath, that can be useful for these cases. pugixml implements an almost complete subset of XPath 1.0. Because of differences in document object model and some performance implications, there are minor violations of the official specifications, which can be found in Conformance to W3C specification. The rest of this section describes the interface for XPath functionality. Please note that if you wish to learn to use XPath language, you have to look for other tutorials or manuals; for example, you can read W3Schools XPath tutorial, XPath tutorial at tizag.com, and the XPath 1.0 specification.

Each XPath expression can have one of the following types: boolean, number, string or node set. Boolean type corresponds to bool type, number type corresponds to double type, string type corresponds to either std::string or std::wstring, depending on whether wide character interface is enabled, and node set corresponds to xpath_node_set type. There is an enumeration, xpath_value_type, which can take the values xpath_type_boolean, xpath_type_number, xpath_type_string or xpath_type_node_set, accordingly.

Because an XPath node can be either a node or an attribute, there is a special type, xpath_node, which is a discriminated union of these types. A value of this type contains two node handles, one of xml_node type, and another one of xml_attribute type; at most one of them can be non-null. The accessors to get these handles are available:

xml_node xpath_node::node() const;
xml_attribute xpath_node::attribute() const;

XPath nodes can be null, in which case both accessors return null handles.

Note that as per XPath specification, each XPath node has a parent, which can be retrieved via this function:

xml_node xpath_node::parent() const;

parent function returns the node's parent if the XPath node corresponds to xml_node handle (equivalent to node().parent()), or the node to which the attribute belongs to, if the XPath node corresponds to xml_attribute handle. For null nodes, parent returns null handle.

Like node and attribute handles, XPath node handles can be implicitly cast to boolean-like object to check if it is a null node, and also can be compared for equality with each other.

You can also create XPath nodes with one of the three constructors: the default constructor, the constructor that takes node argument, and the constructor that takes attribute and node arguments (in which case the attribute must belong to the attribute list of the node). The constructor from xml_node is implicit, so you can usually pass xml_node to functions that expect xpath_node. Apart from that you usually don't need to create your own XPath node objects, since they are returned to you via selection functions.

XPath expressions operate not on single nodes, but instead on node sets. A node set is a collection of nodes, which can be optionally ordered in either a forward document order or a reverse one. Document order is defined in XPath specification; an XPath node is before another node in document order if it appears before it in XML representation of the corresponding document.

Node sets are represented by xpath_node_set object, which has an interface that resembles one of sequential random-access containers. It has an iterator type along with usual begin/past-the-end iterator accessors:

typedef const xpath_node* xpath_node_set::const_iterator;
const_iterator xpath_node_set::begin() const;
const_iterator xpath_node_set::end() const;

And it also can be iterated via indices, just like std::vector:

const xpath_node& xpath_node_set::operator[](size_t index) const;
size_t xpath_node_set::size() const;
bool xpath_node_set::empty() const;

All of the above operations have the same semantics as that of std::vector: the iterators are random-access, all of the above operations are constant time, and accessing the element at index that is greater or equal than the set size results in undefined behavior. You can use both iterator-based and index-based access for iteration, however the iterator-based one can be faster.

The order of iteration depends on the order of nodes inside the set; the order can be queried via the following function:

enum xpath_node_set::type_t {type_unsorted, type_sorted, type_sorted_reverse};
type_t xpath_node_set::type() const;

type function returns the current order of nodes; type_sorted means that the nodes are in forward document order, type_sorted_reverse means that the nodes are in reverse document order, and type_unsorted means that neither order is guaranteed (nodes can accidentally be in a sorted order even if type() returns type_unsorted). If you require a specific order of iteration, you can change it via sort function:

void xpath_node_set::sort(bool reverse = false);

Calling sort sorts the nodes in either forward or reverse document order, depending on the argument; after this call type() will return type_sorted or type_sorted_reverse.

Often the actual iteration is not needed; instead, only the first element in document order is required. For this, a special accessor is provided:

xpath_node xpath_node_set::first() const;

This function returns the first node in forward document order from the set, or null node if the set is empty. Note that while the result of the node does not depend on the order of nodes in the set (i.e. on the result of type()), the complexity does - if the set is sorted, the complexity is constant, otherwise it is linear in the number of elements or worse.

While in the majority of cases the node set is returned by XPath functions, sometimes there is a need to manually construct a node set. For such cases, a constructor is provided which takes an iterator range (const_iterator is a typedef for const xpath_node*), and an optional type:

xpath_node_set::xpath_node_set(const_iterator begin, const_iterator end, type_t type = type_unsorted);

The constructor copies the specified range and sets the specified type. The objects in the range are not checked in any way; you'll have to ensure that the range contains no duplicates, and that the objects are sorted according to the type parameter. Otherwise XPath operations with this set may produce unexpected results.

If you want to select nodes that match some XPath expression, you can do it with the following functions:

xpath_node xml_node::select_single_node(const char_t* query, xpath_variable_set* variables = 0) const;
xpath_node_set xml_node::select_nodes(const char_t* query, xpath_variable_set* variables = 0) const;

select_nodes function compiles the expression and then executes it with the node as a context node, and returns the resulting node set. select_single_node returns only the first node in document order from the result, and is equivalent to calling select_nodes(query).first(). If the XPath expression does not match anything, or the node handle is null, select_nodes returns an empty set, and select_single_node returns null XPath node.

If exception handling is not disabled, both functions throw xpath_exception if the query can not be compiled or if it returns a value with type other than node set; see Error handling for details.

While compiling expressions is fast, the compilation time can introduce a significant overhead if the same expression is used many times on small subtrees. If you're doing many similar queries, consider compiling them into query objects (see Using query objects for further reference). Once you get a compiled query object, you can pass it to select functions instead of an expression string:

xpath_node xml_node::select_single_node(const xpath_query& query) const;
xpath_node_set xml_node::select_nodes(const xpath_query& query) const;

If exception handling is not disabled, both functions throw xpath_exception if the query returns a value with type other than node set.

This is an example of selecting nodes using XPath expressions (samples/xpath_select.cpp):

pugi::xpath_node_set tools = doc.select_nodes("/Profile/Tools/Tool[@AllowRemote='true' and @DeriveCaptionFrom='lastparam']");

std::cout << "Tools:\n";

for (pugi::xpath_node_set::const_iterator it = tools.begin(); it != tools.end(); ++it)
    pugi::xpath_node node = *it;
    std::cout << node.node().attribute("Filename").value() << "\n";

pugi::xpath_node build_tool = doc.select_single_node("//Tool[contains(Description, 'build system')]");

if (build_tool)
    std::cout << "Build tool: " << build_tool.node().attribute("Filename").value() << "\n";

When you call select_nodes with an expression string as an argument, a query object is created behind the scenes. A query object represents a compiled XPath expression. Query objects can be needed in the following circumstances:

  • You can precompile expressions to query objects to save compilation time if it becomes an issue;
  • You can use query objects to evaluate XPath expressions which result in booleans, numbers or strings;
  • You can get the type of expression value via query object.

Query objects correspond to xpath_query type. They are immutable and non-copyable: they are bound to the expression at creation time and can not be cloned. If you want to put query objects in a container, allocate them on heap via new operator and store pointers to xpath_query in the container.

You can create a query object with the constructor that takes XPath expression as an argument:

explicit xpath_query::xpath_query(const char_t* query, xpath_variable_set* variables = 0);

The expression is compiled and the compiled representation is stored in the new query object. If compilation fails, xpath_exception is thrown if exception handling is not disabled (see Error handling for details). After the query is created, you can query the type of the evaluation result using the following function:

xpath_value_type xpath_query::return_type() const;

You can evaluate the query using one of the following functions:

bool xpath_query::evaluate_boolean(const xpath_node& n) const;
double xpath_query::evaluate_number(const xpath_node& n) const;
string_t xpath_query::evaluate_string(const xpath_node& n) const;
xpath_node_set xpath_query::evaluate_node_set(const xpath_node& n) const;

All functions take the context node as an argument, compute the expression and return the result, converted to the requested type. According to XPath specification, value of any type can be converted to boolean, number or string value, but no type other than node set can be converted to node set. Because of this, evaluate_boolean, evaluate_number and evaluate_string always return a result, but evaluate_node_set results in an error if the return type is not node set (see Error handling).

[Note] Note

Calling node.select_nodes("query") is equivalent to calling xpath_query("query").evaluate_node_set(node).

Note that evaluate_string function returns the STL string; as such, it's not available in PUGIXML_NO_STL mode and also usually allocates memory. There is another string evaluation function:

size_t xpath_query::evaluate_string(char_t* buffer, size_t capacity, const xpath_node& n) const;

This function evaluates the string, and then writes the result to buffer (but at most capacity characters); then it returns the full size of the result in characters, including the terminating zero. If capacity is not 0, the resulting buffer is always zero-terminated. You can use this function as follows:

  • First call the function with buffer = 0 and capacity = 0; then allocate the returned amount of characters, and call the function again, passing the allocated storage and the amount of characters;
  • First call the function with small buffer and buffer capacity; then, if the result is larger than the capacity, the output has been trimmed, so allocate a larger buffer and call the function again.

This is an example of using query objects (samples/xpath_query.cpp):

// Select nodes via compiled query
pugi::xpath_query query_remote_tools("/Profile/Tools/Tool[@AllowRemote='true']");

pugi::xpath_node_set tools = query_remote_tools.evaluate_node_set(doc);
std::cout << "Remote tool: ";

// Evaluate numbers via compiled query
pugi::xpath_query query_timeouts("sum(//Tool/@Timeout)");
std::cout << query_timeouts.evaluate_number(doc) << std::endl;

// Evaluate strings via compiled query for different context nodes
pugi::xpath_query query_name_valid("string-length(substring-before(@Filename, '_')) > 0 and @OutputFileMasks");
pugi::xpath_query query_name("concat(substring-before(@Filename, '_'), ' produces ', @OutputFileMasks)");

for (pugi::xml_node tool = doc.first_element_by_path("Profile/Tools/Tool"); tool; tool = tool.next_sibling())
    std::string s = query_name.evaluate_string(tool);

    if (query_name_valid.evaluate_boolean(tool)) std::cout << s << std::endl;

XPath queries may contain references to variables; this is useful if you want to use queries that depend on some dynamic parameter without manually preparing the complete query string, or if you want to reuse the same query object for similar queries.

Variable references have the form $name; in order to use them, you have to provide a variable set, which includes all variables present in the query with correct types. This set is passed to xpath_query constructor or to select_nodes/select_single_node functions:

explicit xpath_query::xpath_query(const char_t* query, xpath_variable_set* variables = 0);
xpath_node xml_node::select_single_node(const char_t* query, xpath_variable_set* variables = 0) const;
xpath_node_set xml_node::select_nodes(const char_t* query, xpath_variable_set* variables = 0) const;

If you're using query objects, you can change the variable values before evaluate/select calls to change the query behavior.

[Note] Note

The variable set pointer is stored in the query object; you have to ensure that the lifetime of the set exceeds that of query object.

Variable sets correspond to xpath_variable_set type, which is essentially a variable container.

You can add new variables with the following function:

xpath_variable* xpath_variable_set::add(const char_t* name, xpath_value_type type);

The function tries to add a new variable with the specified name and type; if the variable with such name does not exist in the set, the function adds a new variable and returns the variable handle; if there is already a variable with the specified name, the function returns the variable handle if variable has the specified type. Otherwise the function returns null pointer; it also returns null pointer on allocation failure.

New variables are assigned the default value which depends on the type: 0 for numbers, false for booleans, empty string for strings and empty set for node sets.

You can get the existing variables with the following functions:

xpath_variable* xpath_variable_set::get(const char_t* name);
const xpath_variable* xpath_variable_set::get(const char_t* name) const;

The functions return the variable handle, or null pointer if the variable with the specified name is not found.

Additionally, there are the helper functions for setting the variable value by name; they try to add the variable with the corresponding type, if it does not exist, and to set the value. If the variable with the same name but with different type is already present, they return false; they also return false on allocation failure. Note that these functions do not perform any type conversions.

bool xpath_variable_set::set(const char_t* name, bool value);
bool xpath_variable_set::set(const char_t* name, double value);
bool xpath_variable_set::set(const char_t* name, const char_t* value);
bool xpath_variable_set::set(const char_t* name, const xpath_node_set& value);

The variable values are copied to the internal variable storage, so you can modify or destroy them after the functions return.

If setting variables by name is not efficient enough, or if you have to inspect variable information or get variable values, you can use variable handles. A variable corresponds to the xpath_variable type, and a variable handle is simply a pointer to xpath_variable.

In order to get variable information, you can use one of the following functions:

const char_t* xpath_variable::name() const;
xpath_value_type xpath_variable::type() const;

Note that each variable has a distinct type which is specified upon variable creation and can not be changed later.

In order to get variable value, you should use one of the following functions, depending on the variable type:

bool xpath_variable::get_boolean() const;
double xpath_variable::get_number() const;
const char_t* xpath_variable::get_string() const;
const xpath_node_set& xpath_variable::get_node_set() const;

These functions return the value of the variable. Note that no type conversions are performed; if the type mismatch occurs, a dummy value is returned (false for booleans, NaN for numbers, empty string for strings and empty set for node sets).

In order to set variable value, you should use one of the following functions, depending on the variable type:

bool xpath_variable::set(bool value);
bool xpath_variable::set(double value);
bool xpath_variable::set(const char_t* value);
bool xpath_variable::set(const xpath_node_set& value);

These functions modify the variable value. Note that no type conversions are performed; if the type mismatch occurs, the functions return false; they also return false on allocation failure. The variable values are copied to the internal variable storage, so you can modify or destroy them after the functions return.

This is an example of using variables in XPath queries (samples/xpath_variables.cpp):

// Select nodes via compiled query
pugi::xpath_variable_set vars;
vars.add("remote", pugi::xpath_type_boolean);

pugi::xpath_query query_remote_tools("/Profile/Tools/Tool[@AllowRemote = string($remote)]", &vars);

vars.set("remote", true);
pugi::xpath_node_set tools_remote = query_remote_tools.evaluate_node_set(doc);

vars.set("remote", false);
pugi::xpath_node_set tools_local = query_remote_tools.evaluate_node_set(doc);

std::cout << "Remote tool: ";

std::cout << "Local tool: ";

// You can pass the context directly to select_nodes/select_single_node
pugi::xpath_node_set tools_local_imm = doc.select_nodes("/Profile/Tools/Tool[@AllowRemote = string($remote)]", &vars);

std::cout << "Local tool imm: ";

There are two different mechanisms for error handling in XPath implementation; the mechanism used depends on whether exception support is disabled (this is controlled with PUGIXML_NO_EXCEPTIONS define).

By default, XPath functions throw xpath_exception object in case of errors; additionally, in the event any memory allocation fails, an std::bad_alloc exception is thrown. Also xpath_exception is thrown if the query is evaluated to a node set, but the return type is not node set. If the query constructor succeeds (i.e. no exception is thrown), the query object is valid. Otherwise you can get the error details via one of the following functions:

virtual const char* xpath_exception::what() const throw();
const xpath_parse_result& xpath_exception::result() const;

If exceptions are disabled, then in the event of parsing failure the query is initialized to invalid state; you can test if the query object is valid by using it in a boolean expression: if (query) { ... }. Additionally, you can get parsing result via the result() accessor:

const xpath_parse_result& xpath_query::result() const;

Without exceptions, evaluating invalid query results in false, empty string, NaN or an empty node set, depending on the type; evaluating a query as a node set results in an empty node set if the return type is not node set.

The information about parsing result is returned via xpath_parse_result object. It contains parsing status and the offset of last successfully parsed character from the beginning of the source stream:

struct xpath_parse_result
    const char* error;
    ptrdiff_t offset;

    operator bool() const;
    const char* description() const;

Parsing result is represented as the error message; it is either a null pointer, in case there is no error, or the error message in the form of ASCII zero-terminated string.

description() member function can be used to get the error message; it never returns the null pointer, so you can safely use description() even if query parsing succeeded. Note that description() returns a char string even in PUGIXML_WCHAR_MODE; you'll have to call as_wide to get the wchar_t string.

In addition to the error message, parsing result has an offset member, which contains the offset of last successfully parsed character. This offset is in units of pugi::char_t (bytes for character mode, wide characters for wide character mode).

Parsing result object can be implicitly converted to bool like this: if (result) { ... } else { ... }.

This is an example of XPath error handling (samples/xpath_error.cpp):

// Exception is thrown for incorrect query syntax
catch (const pugi::xpath_exception& e)
    std::cout << "Select failed: " << e.what() << std::endl;

// Exception is thrown for incorrect query semantics
catch (const pugi::xpath_exception& e)
    std::cout << "Select failed: " << e.what() << std::endl;

// Exception is thrown for query with incorrect return type
catch (const pugi::xpath_exception& e)
    std::cout << "Select failed: " << e.what() << std::endl;

Because of the differences in document object models, performance considerations and implementation complexity, pugixml does not provide a fully conformant XPath 1.0 implementation. This is the current list of incompatibilities:

  • Consecutive text nodes sharing the same parent are not merged, i.e. in <node>text1 <![CDATA[data]]> text2</node> node should have one text node child, but instead has three.
  • Since the document type declaration is not used for parsing, id() function always returns an empty node set.
  • Namespace nodes are not supported (affects namespace:: axis).
  • Name tests are performed on QNames in XML document instead of expanded names; for <foo xmlns:ns1='uri' xmlns:ns2='uri'><ns1:child/><ns2:child/></foo>, query foo/ns1:* will return only the first child, not both of them. Compliant XPath implementations can return both nodes if the user provides appropriate namespace declarations.
  • String functions consider a character to be either a single char value or a single wchar_t value, depending on the library configuration; this means that some string functions are not fully Unicode-aware. This affects substring(), string-length() and translate() functions.

pugixml 1.4 manual | Overview | Installation | Document: Object model · Loading · Accessing · Modifying · Saving | XPath | API Reference | Table of Contents