xmllint match and extract value

Question

I would like to pull only "image" from the xml but I'm getting an error since the first couple ones don't include "image".

<?xml version='1.0' encoding='utf-8'?>
<document>
    <job name="Job1">
        <type>
            <description>
            </description>
        </type>
    </job>
    <job name="Job2">
        <type>
            <description>
            </description>
        </type>
    </job>
    <job name="Job3">
        <type>
            <description>
                <image>
                    <png></png>
                </image>
            </description>
        </type>
    </job>
</document>

How can it skip the first two and match the third image tag?

xmllint --xpath "//*[local-name()='document']/job/type/description/image/png/text()" file

Answer 1

If I modify only one line of your input file, changing

<png></png>

to

<png>Some text goes here</png>

...then your code works perfectly as already written. Thus, the error has nothing at all to do with the first few descriptions having no png.

---

Mind you, since you don't have any namespaces defined, you don't need to mess with local-name() at all.

xmllint --xpath "/document/job/type/description/image/png/text()"

...works just as well.