Question

I’m trying to convert an xml document to Python data structures.

A sample of the xml:

    <SOFTWARES>
        <PUBLISHER>Microsoft Corporation</PUBLISHER>
        <NAME>Microsoft Office Visio 2010</NAME>
        <VERSION>14.0.6029.1000</VERSION>
        <FOLDER>C:/Program Files/Microsoft Office/</FOLDER>
        <LANGUAGE>Language Neutral</LANGUAGE>
        <INSTALLDATE>2012/03/29</INSTALLDATE>
    </SOFTWARES>
    <SOFTWARES>
        <PUBLISHER>Microsoft</PUBLISHER>
        <NAME>Update for Microsoft Office 2010 (KB2553310) 64-Bit Edition</NAME>
        <INSTALLDATE>0000//0/0/00</INSTALLDATE>
    </SOFTWARES>

lxml.de has an excellent example of this: http://lxml.de/FAQ.html#how-can-i-map-an-xml-tree-into-a-dict-of-dicts

def xml_to_dict(element):
    return element.tag, dict(map(xml_to_dict, element)) or element.text

This produces a great dict of dicts which has but one flaw. It will override existing keys. So when the the process is complete I get:

'SOFTWARES': {
    'PUBLISHER': 'Microsoft',
    'NAME': 'Update for Microsoft Office 2010 (KB2553310) 64-Bit Edition',
    'INSTALLDATE': '0000//0/0/00',
},

Which is the last SOFTWARES block, regardless of how many were before it. lxml’s function works well because it’s recursive but I want to write something that can handle duplicate keys. Preferably by just tossing the SOFTWARES dicts in a list and I can just iterate through the list when the time comes.

Answer 1

Easiest solution for this specific case:

map(xml_to_dict, xml_document)

This will give you a list of dictionaries.