Skip to content

Latest commit

 

History

History
117 lines (90 loc) · 5.04 KB

reading-a-document.md

File metadata and controls

117 lines (90 loc) · 5.04 KB
Raw

Reading a Document

When reading a XML document with libxmlmm there are two options. You can either navigate the document by querying each element for it's children or you can use XPATH queries into the document.

For this example we have the following simple XML:

<?xml version='1.0'?>
<message version="1.2">
  <from>Mack</from>
  <to>Joe</to>
  <to>Sally</to>
  <to>Mike</to>
  <body>Hello everybody!</body>
</message>

It is our goal to get all recipients; the names that are in the to elements.

Navigating the Document

To navigate the document you basically use 'get_children' all the time.

xml::Document doc;
doc.read_from_file("message.xml");

xml::Element* message = doc.get_root_node();
assert(message != NULL);
assert(message->get_name() == "message");
assert(message->get_attribute("version") == "1.2");

std::vector<xml::Node*> children = root->get_children();
std::vector<std::string> recipients;
for (unsigned int i; i < children.size(); i++)
{
    xml::Element* element = dynamic_cast<xml::Element*>(children[i]);
    if (element != NULL && element->get_name() == "to")
    {
        recipients.push_back(element->get_text());
    }
}

Basically you create a document; in this case we simply create it on the stack. This has the advantage that we don't have to bother about it being deleted. The same is with all child objects, they are managed by their parents and as a result by the document. The important bit to remember is that a node will be deleted once the document is deleted. So if you happen to pass node pointers around, you need to ensure that your document lives longer than the context in which the pointer is used.

After creating the document you can read it from file with read_from_file. Alternately you can also read it from any stream with read_from_stream.

The first thing you need to do is get the root element. This is simply done by calling 'get_root_element'. In this case we get an Element object. In the case here we check if the root node really is a message element and the attribute 'version' is 1.2.

To access the children of the root element, you call 'get_children'. Here you get a vector of Node object. These can be anything from Element, TextNode to CDataNode. In contrast to other DOM implementations you do not get attributes as Node object. They can only be accessed though get_attribute and friends. To check the type of node, you need to use dynamic_cast. In this case we are looking for nodes of type Element that are "to".

As a simplification, the inner text of an element can be accessed though get_text. It would also be possible to call get_children and cast the resulting Node to TextNode and call get_value on it.

XPath

libxmlmm features the ability to use XPath to access values in a document. To access single values you can use query_string and query_number and to access nodes you can use find_node for one node and find_nodes for multiple nodes. As a shortcut you may also use find_element and find_elements if you are sure that the result of your query are only elements.

xml::Document doc;
doc.read_from_file("message.xml");

assert("1.2" == doc.query_string("/message/@version"));

double recipients_count = doc.query_number("count(/message/to)");
std::vector<std::string> recipients(static_cast<size_t>(recipients_count));

std::vector<xml::Element*> to_elements doc.find_elements("/message/to");
assert(recipients_count == to_elements.size());
for (unsigned int i; i < to_elements.size(); i++)
{
    recipients.push_back(to_elements[i]->get_text());
}

This approach starts like the other with creating and reading a document.

First we check the version by calling query_string. The important thing to remember is that when the XPath query fails it will throw a InvalidXPath exception, so by simply calling query_string, it is ensured that there is a root element named 'message' and it has an attribute names 'version'.

The next thing we do is demonstrate query_number. The clever programmer will notice that we can also get the value from to_elements.size() and that is probably 100 times faster. Nevertheless query_number will return a double containing the number of 'to' elements. 'query_number' is especially useful for XPath queries that do not return nodes, such as the XPath function count.

Finally we query all 'to' elements with query_elements. This can also be done with query_nodes and the code would look like the enumeration example. This shows why this method is so useful, it make much cleaner code in those cases where you are accessing only elements; and experience shows that when working with XML you will probably query 80% of the time for elements.

Conclusion

Which approach you take depends on your use case. Basically the XPath approach is simpler, but will probably be slower, since the XPath must be evaluated. It is up to you to gauge the tradeoff between speed and simplicity.