XSLT, 2nd Edition

Advantages of the key() Function

Now that we’ve taken the key() function through its paces, you can see that it has several advantages:

• The key() function is defined in a stylesheet. That means I can define any number of relationships between parts of an XML document at any time. If I need to define a new relationship tomorrow, I don’t have to change my XML documents.

• Any number of key() functions can be defined for a given element. In our parts list example, we could define key() functions for the values of the vendor-id, part-id, and component-id attributes. We could also create key() functions based on the text of various elements or their children. If we used IDs instead of the key() function, we would be limited to a single index based on the value of the single attribute of the ID datatype.

  To sum up the advantages for this point, an element can have more than one key() defined against it, and that key doesn’t have to be based on an attribute. The key can be based on the element’s text, the text of child elements, or other constructs.

• Any number of elements can match a given value. Taking another look at our example, when we use the key() function to find all the parts from a particular country, the key() function returns a node-set that can have any number of nodes. If we use an ID instead, legally there can be only one element that matches a given country.

• The value we use to look up elements in the key function isn’t constrained to be an XML name. If we use the ID datatype, ...