I like XPath! The w3schools tutorial is OK.

xmllint xmltest.xml --xpath '//Unit[UnitType="Type2"]/SerialNumber/text()'

will do it. It's a bit 'lazy' as it will find all 'Unit' elements with a child of 'UnitType' 'Type2' even if they are used elsewhere in the document, but if that's not a problem keep it simple. The text() takes off the <SerialNumber> element name.

The xmllint command comes from libxml2 package.

Cheers Neil

On 07/11/2013 10:35, Mark Rogers wrote:

Given the following:

<?xml version="1.0" encoding="utf-8"?>

<Project> <Network> <Unit> <UnitType>Type1</UnitType> <SerialNumber>123456</SerialNumber> </Unit> <Unit> <UnitType>Type2</UnitType> <SerialNumber>abcdef</SerialNumber> </Unit> [... snip...] </Network> </Project>

.. I need to extract all the serial numbers where UnitType is "Type2".

I have never managed to get my head around XPath etc, and all the examples I find seem either too general or too complex for me to make much progress. I always end up writing some code that parses it as text which is a massive cop out.

How best can I do this correctly using XML tools, and which tools should I use?

Mark