How to select multiple nodes in different levels?

Question

Having this (simplified) XML:

<?xml version="1.0" encoding="UTF-8"?>
<kml>
<Document>
        <Placemark>
            <name>Poly 1</name>
            <Polygon>
                        <coordinates>
                            -58.40844625779582,-34.60295278618136,0
                        </coordinates>
            </Polygon>
        </Placemark>
        <Placemark>
            <name>Poly 2</name>
            <Polygon>
                        <coordinates>
                            -58.40414334150432,-34.59992445476809,0
                        </coordinates>
            </Polygon>
        </Placemark>
</Document>
</kml>

How can I select the name and coordinates of each Placemark? Right now I can select their name with the following XPath expression:

//Document//Placemark//name

How can I select both without any other data?

Solution 1

You can use a union in your XPath expression. Just use the operator: |

//Document/Placemark/name | //Document/Placemark/Polygon/coordinates

Don't use the // (descendant axis) if you don't need to. Using //, this would also work: //name | //coordinates. It's better performance-wise to specify the exact path.

Solution 2

Use:

/*/*/Placemark/name | /*/*/Placemark/*/coordinates

This specifies the union of the results of two separate XPath expressions -- the standard XPath union operator | is used. Selected are all name elements that are children of a Placemark element that is a grandchild of the top element of the XML document, plus all coordinates elements that are grand-children of a Placemark element that is a grandchild of the top element of the XML document.

The selected elements come in document order (although no normative W3C document specifies the order), which means that in the result of the evaluation (usually of type XmlNodeList) any name element is directly followed by its corresponding coordinates element.

Comments

[U. Windl]

Is there a simplification that avoids repeating //Document/Placemark/?