[python] How to convert an xml string to a dictionary?

I have a program that reads an xml document from a socket. I have the xml document stored in a string which I would like to convert directly to a Python dictionary, the same way it is done in Django's simplejson library.

Take as an example:

str ="<?xml version="1.0" ?><person><name>john</name><age>20</age></person"
dic_xml = convert_to_dic(str)

Then dic_xml would look like {'person' : { 'name' : 'john', 'age' : 20 } }

@dibrovsd: Solution will not work if the xml have more than one tag with same name

On your line of thought, I have modified the code a bit and written it for general node instead of root:

from collections import defaultdict
def xml2dict(node):
    d, count = defaultdict(list), 1
    for i in node:
        d[i.tag + "_" + str(count)]['text'] = i.findtext('.')[0]
        d[i.tag + "_" + str(count)]['attrib'] = i.attrib # attrib gives the list
        d[i.tag + "_" + str(count)]['children'] = xml2dict(i) # it gives dict
     return d

The code from http://code.activestate.com/recipes/410469-xml-as-dictionary/ works well, but if there are multiple elements that are the same at a given place in the hierarchy it just overrides them.

I added a shim between that looks to see if the element already exists before self.update(). If so, pops the existing entry and creates a lists out of the existing and the new. Any subsequent duplicates are added to the list.

Not sure if this can be handled more gracefully, but it works:

import xml.etree.ElementTree as ElementTree

class XmlDictConfig(dict):
    def __init__(self, parent_element):
        if parent_element.items():
            self.updateShim(dict(parent_element.items()))
        for element in parent_element:
            if len(element):
                aDict = XmlDictConfig(element)
                if element.items():
                    aDict.updateShim(dict(element.items()))
                self.updateShim({element.tag: aDict})
            elif element.items():
                self.updateShim({element.tag: dict(element.items())})
            else:
                self.updateShim({element.tag: element.text.strip()})

    def updateShim (self, aDict ):
        for key in aDict.keys():
            if key in self:
                value = self.pop(key)
                if type(value) is not list:
                    listOfDicts = []
                    listOfDicts.append(value)
                    listOfDicts.append(aDict[key])
                    self.update({key: listOfDicts})

                else:
                    value.append(aDict[key])
                    self.update({key: value})
            else:
                self.update(aDict)

The following XML-to-Python-dict snippet parses entities as well as attributes following this XML-to-JSON "specification". It is the most general solution handling all cases of XML.

from collections import defaultdict

def etree_to_dict(t):
    d = {t.tag: {} if t.attrib else None}
    children = list(t)
    if children:
        dd = defaultdict(list)
        for dc in map(etree_to_dict, children):
            for k, v in dc.items():
                dd[k].append(v)
        d = {t.tag: {k:v[0] if len(v) == 1 else v for k, v in dd.items()}}
    if t.attrib:
        d[t.tag].update(('@' + k, v) for k, v in t.attrib.items())
    if t.text:
        text = t.text.strip()
        if children or t.attrib:
            if text:
              d[t.tag]['#text'] = text
        else:
            d[t.tag] = text
    return d

It is used:

from xml.etree import cElementTree as ET
e = ET.XML('''
<root>
  <e />
  <e>text</e>
  <e name="value" />
  <e name="value">text</e>
  <e> <a>text</a> <b>text</b> </e>
  <e> <a>text</a> <a>text</a> </e>
  <e> text <a>text</a> </e>
</root>
''')

from pprint import pprint
pprint(etree_to_dict(e))

The output of this example (as per above-linked "specification") should be:

{'root': {'e': [None,
                'text',
                {'@name': 'value'},
                {'#text': 'text', '@name': 'value'},
                {'a': 'text', 'b': 'text'},
                {'a': ['text', 'text']},
                {'#text': 'text', 'a': 'text'}]}}

Not necessarily pretty, but it is unambiguous, and simpler XML inputs result in simpler JSON. :)

Update

If you want to do the reverse, emit an XML string from a JSON/dict, you can use:

try:
  basestring
except NameError:  # python3
  basestring = str

def dict_to_etree(d):
    def _to_etree(d, root):
        if not d:
            pass
        elif isinstance(d, basestring):
            root.text = d
        elif isinstance(d, dict):
            for k,v in d.items():
                assert isinstance(k, basestring)
                if k.startswith('#'):
                    assert k == '#text' and isinstance(v, basestring)
                    root.text = v
                elif k.startswith('@'):
                    assert isinstance(v, basestring)
                    root.set(k[1:], v)
                elif isinstance(v, list):
                    for e in v:
                        _to_etree(e, ET.SubElement(root, k))
                else:
                    _to_etree(v, ET.SubElement(root, k))
        else:
            raise TypeError('invalid type: ' + str(type(d)))
    assert isinstance(d, dict) and len(d) == 1
    tag, body = next(iter(d.items()))
    node = ET.Element(tag)
    _to_etree(body, node)
    return ET.tostring(node)

pprint(dict_to_etree(d))

xmltodict (full disclosure: I wrote it) does exactly that:

xmltodict.parse("""
<?xml version="1.0" ?>
<person>
  <name>john</name>
  <age>20</age>
</person>""")
# {u'person': {u'age': u'20', u'name': u'john'}}

I wrote a simple recursive function to do the job:

from xml.etree import ElementTree
root = ElementTree.XML(xml_to_convert)

def xml_to_dict_recursive(root):

    if len(root.getchildren()) == 0:
        return {root.tag:root.text}
    else:
        return {root.tag:list(map(xml_to_dict_recursive, root.getchildren()))}

The easiest to use XML parser for Python is ElementTree (as of 2.5x and above it is in the standard library xml.etree.ElementTree). I don't think there is anything that does exactly what you want out of the box. It would be pretty trivial to write something to do what you want using ElementTree, but why convert to a dictionary, and why not just use ElementTree directly.

I have modified one of the answers to my taste and to work with multiple values with the same tag for example consider the following xml code saved in XML.xml file

     <A>
        <B>
            <BB>inAB</BB>
            <C>
                <D>
                    <E>
                        inABCDE
                    </E>
                    <E>value2</E>
                    <E>value3</E>
                </D>
                <inCout-ofD>123</inCout-ofD>
            </C>
        </B>
        <B>abc</B>
        <F>F</F>
    </A>

and in python

import xml.etree.ElementTree as ET




class XMLToDictionary(dict):
    def __init__(self, parentElement):
        self.parentElement = parentElement
        for child in list(parentElement):
            child.text = child.text if (child.text != None) else  ' '
            if len(child) == 0:
                self.update(self._addToDict(key= child.tag, value = child.text.strip(), dict = self))
            else:
                innerChild = XMLToDictionary(parentElement=child)
                self.update(self._addToDict(key=innerChild.parentElement.tag, value=innerChild, dict=self))

    def getDict(self):
        return {self.parentElement.tag: self}

    class _addToDict(dict):
        def __init__(self, key, value, dict):
            if not key in dict:
                self.update({key: value})
            else:
                identical = dict[key] if type(dict[key]) == list else [dict[key]]
                self.update({key: identical + [value]})


tree = ET.parse('./XML.xml')
root = tree.getroot()
parseredDict = XMLToDictionary(root).getDict()
print(parseredDict)

the output is

{'A': {'B': [{'BB': 'inAB', 'C': {'D': {'E': ['inABCDE', 'value2', 'value3']}, 'inCout-ofD': '123'}}, 'abc'], 'F': 'F'}}

Here's a link to an ActiveState solution - and the code in case it disappears again.

==================================================
xmlreader.py:
==================================================
from xml.dom.minidom import parse


class NotTextNodeError:
    pass


def getTextFromNode(node):
    """
    scans through all children of node and gathers the
    text. if node has non-text child-nodes, then
    NotTextNodeError is raised.
    """
    t = ""
    for n in node.childNodes:
    if n.nodeType == n.TEXT_NODE:
        t += n.nodeValue
    else:
        raise NotTextNodeError
    return t


def nodeToDic(node):
    """
    nodeToDic() scans through the children of node and makes a
    dictionary from the content.
    three cases are differentiated:
    - if the node contains no other nodes, it is a text-node
    and {nodeName:text} is merged into the dictionary.
    - if the node has the attribute "method" set to "true",
    then it's children will be appended to a list and this
    list is merged to the dictionary in the form: {nodeName:list}.
    - else, nodeToDic() will call itself recursively on
    the nodes children (merging {nodeName:nodeToDic()} to
    the dictionary).
    """
    dic = {} 
    for n in node.childNodes:
    if n.nodeType != n.ELEMENT_NODE:
        continue
    if n.getAttribute("multiple") == "true":
        # node with multiple children:
        # put them in a list
        l = []
        for c in n.childNodes:
            if c.nodeType != n.ELEMENT_NODE:
            continue
        l.append(nodeToDic(c))
            dic.update({n.nodeName:l})
        continue

    try:
        text = getTextFromNode(n)
    except NotTextNodeError:
            # 'normal' node
            dic.update({n.nodeName:nodeToDic(n)})
            continue

        # text node
        dic.update({n.nodeName:text})
    continue
    return dic


def readConfig(filename):
    dom = parse(filename)
    return nodeToDic(dom)





def test():
    dic = readConfig("sample.xml")

    print dic["Config"]["Name"]
    print
    for item in dic["Config"]["Items"]:
    print "Item's Name:", item["Name"]
    print "Item's Value:", item["Value"]

test()



==================================================
sample.xml:
==================================================
<?xml version="1.0" encoding="UTF-8"?>

<Config>
    <Name>My Config File</Name>

    <Items multiple="true">
    <Item>
        <Name>First Item</Name>
        <Value>Value 1</Value>
    </Item>
    <Item>
        <Name>Second Item</Name>
        <Value>Value 2</Value>
    </Item>
    </Items>

</Config>



==================================================
output:
==================================================
My Config File

Item's Name: First Item
Item's Value: Value 1
Item's Name: Second Item
Item's Value: Value 2

def xml_to_dict(node):
    u''' 
    @param node:lxml_node
    @return: dict 
    '''

    return {'tag': node.tag, 'text': node.text, 'attrib': node.attrib, 'children': {child.tag: xml_to_dict(child) for child in node}}

At one point I had to parse and write XML that only consisted of elements without attributes so a 1:1 mapping from XML to dict was possible easily. This is what I came up with in case someone else also doesnt need attributes:

def xmltodict(element):
    if not isinstance(element, ElementTree.Element):
        raise ValueError("must pass xml.etree.ElementTree.Element object")

    def xmltodict_handler(parent_element):
        result = dict()
        for element in parent_element:
            if len(element):
                obj = xmltodict_handler(element)
            else:
                obj = element.text

            if result.get(element.tag):
                if hasattr(result[element.tag], "append"):
                    result[element.tag].append(obj)
                else:
                    result[element.tag] = [result[element.tag], obj]
            else:
                result[element.tag] = obj
        return result

    return {element.tag: xmltodict_handler(element)}


def dicttoxml(element):
    if not isinstance(element, dict):
        raise ValueError("must pass dict type")
    if len(element) != 1:
        raise ValueError("dict must have exactly one root key")

    def dicttoxml_handler(result, key, value):
        if isinstance(value, list):
            for e in value:
                dicttoxml_handler(result, key, e)
        elif isinstance(value, basestring):
            elem = ElementTree.Element(key)
            elem.text = value
            result.append(elem)
        elif isinstance(value, int) or isinstance(value, float):
            elem = ElementTree.Element(key)
            elem.text = str(value)
            result.append(elem)
        elif value is None:
            result.append(ElementTree.Element(key))
        else:
            res = ElementTree.Element(key)
            for k, v in value.items():
                dicttoxml_handler(res, k, v)
            result.append(res)

    result = ElementTree.Element(element.keys()[0])
    for key, value in element[element.keys()[0]].items():
        dicttoxml_handler(result, key, value)
    return result

def xmlfiletodict(filename):
    return xmltodict(ElementTree.parse(filename).getroot())

def dicttoxmlfile(element, filename):
    ElementTree.ElementTree(dicttoxml(element)).write(filename)

def xmlstringtodict(xmlstring):
    return xmltodict(ElementTree.fromstring(xmlstring).getroot())

def dicttoxmlstring(element):
    return ElementTree.tostring(dicttoxml(element))

This lightweight version, while not configurable, is pretty easy to tailor as needed, and works in old pythons. Also it is rigid - meaning the results are the same regardless of the existence of attributes.

import xml.etree.ElementTree as ET

from copy import copy

def dictify(r,root=True):
    if root:
        return {r.tag : dictify(r, False)}
    d=copy(r.attrib)
    if r.text:
        d["_text"]=r.text
    for x in r.findall("./*"):
        if x.tag not in d:
            d[x.tag]=[]
        d[x.tag].append(dictify(x,False))
    return d

So:

root = ET.fromstring("<erik><a x='1'>v</a><a y='2'>w</a></erik>")

dictify(root)

Results in:

{'erik': {'a': [{'x': '1', '_text': 'v'}, {'y': '2', '_text': 'w'}]}}

Disclaimer: This modified XML parser was inspired by Adam Clark The original XML parser works for most of simple cases. However, it didn't work for some complicated XML files. I debugged the code line by line and finally fixed some issues. If you find some bugs, please let me know. I am glad to fix it.

class XmlDictConfig(dict):  
    '''   
    Note: need to add a root into if no exising    
    Example usage:
    >>> tree = ElementTree.parse('your_file.xml')
    >>> root = tree.getroot()
    >>> xmldict = XmlDictConfig(root)
    Or, if you want to use an XML string:
    >>> root = ElementTree.XML(xml_string)
    >>> xmldict = XmlDictConfig(root)
    And then use xmldict for what it is... a dict.
    '''
    def __init__(self, parent_element):
        if parent_element.items():
            self.updateShim( dict(parent_element.items()) )
        for element in parent_element:
            if len(element):
                aDict = XmlDictConfig(element)
            #   if element.items():
            #   aDict.updateShim(dict(element.items()))
                self.updateShim({element.tag: aDict})
            elif element.items():    # items() is specialy for attribtes
                elementattrib= element.items()
                if element.text:           
                    elementattrib.append((element.tag,element.text ))     # add tag:text if there exist
                self.updateShim({element.tag: dict(elementattrib)})
            else:
                self.updateShim({element.tag: element.text})

    def updateShim (self, aDict ):
        for key in aDict.keys():   # keys() includes tag and attributes
            if key in self:
                value = self.pop(key)
                if type(value) is not list:
                    listOfDicts = []
                    listOfDicts.append(value)
                    listOfDicts.append(aDict[key])
                    self.update({key: listOfDicts})
                else:
                    value.append(aDict[key])
                    self.update({key: value})
            else:
                self.update({key:aDict[key]})  # it was self.update(aDict)    

You can do this quite easily with lxml. First install it:

[sudo] pip install lxml

Here is a recursive function I wrote that does the heavy lifting for you:

from lxml import objectify as xml_objectify


def xml_to_dict(xml_str):
    """ Convert xml to dict, using lxml v3.4.2 xml processing library """
    def xml_to_dict_recursion(xml_object):
        dict_object = xml_object.__dict__
        if not dict_object:
            return xml_object
        for key, value in dict_object.items():
            dict_object[key] = xml_to_dict_recursion(value)
        return dict_object
    return xml_to_dict_recursion(xml_objectify.fromstring(xml_str))

xml_string = """<?xml version="1.0" encoding="UTF-8"?><Response><NewOrderResp>
<IndustryType>Test</IndustryType><SomeData><SomeNestedData1>1234</SomeNestedData1>
<SomeNestedData2>3455</SomeNestedData2></SomeData></NewOrderResp></Response>"""

print xml_to_dict(xml_string)

The below variant preserves the parent key / element:

def xml_to_dict(xml_str):
    """ Convert xml to dict, using lxml v3.4.2 xml processing library, see http://lxml.de/ """
    def xml_to_dict_recursion(xml_object):
        dict_object = xml_object.__dict__
        if not dict_object:  # if empty dict returned
            return xml_object
        for key, value in dict_object.items():
            dict_object[key] = xml_to_dict_recursion(value)
        return dict_object
    xml_obj = objectify.fromstring(xml_str)
    return {xml_obj.tag: xml_to_dict_recursion(xml_obj)}

If you want to only return a subtree and convert it to dict, you can use Element.find() to get the subtree and then convert it:

xml_obj.find('.//')  # lxml.objectify.ObjectifiedElement instance

See the lxml docs here. I hope this helps!

From @K3---rnc response (the best for me) I've added a small modifications to get an OrderedDict from an XML text (some times order matters):

def etree_to_ordereddict(t):
d = OrderedDict()
d[t.tag] = OrderedDict() if t.attrib else None
children = list(t)
if children:
    dd = OrderedDict()
    for dc in map(etree_to_ordereddict, children):
        for k, v in dc.iteritems():
            if k not in dd:
                dd[k] = list()
            dd[k].append(v)
    d = OrderedDict()
    d[t.tag] = OrderedDict()
    for k, v in dd.iteritems():
        if len(v) == 1:
            d[t.tag][k] = v[0]
        else:
            d[t.tag][k] = v
if t.attrib:
    d[t.tag].update(('@' + k, v) for k, v in t.attrib.iteritems())
if t.text:
    text = t.text.strip()
    if children or t.attrib:
        if text:
            d[t.tag]['#text'] = text
    else:
        d[t.tag] = text
return d

Following @K3---rnc example, you can use it:

from xml.etree import cElementTree as ET
e = ET.XML('''
<root>
  <e />
  <e>text</e>
  <e name="value" />
  <e name="value">text</e>
  <e> <a>text</a> <b>text</b> </e>
  <e> <a>text</a> <a>text</a> </e>
  <e> text <a>text</a> </e>
</root>
''')

from pprint import pprint
pprint(etree_to_ordereddict(e))

Hope it helps ;)

The most recent versions of the PicklingTools libraries (1.3.0 and 1.3.1) support tools for converting from XML to a Python dict.

The download is available here: PicklingTools 1.3.1

There is quite a bit of documentation for the converters here: the documentation describes in detail all of the decisions and issues that will arise when converting between XML and Python dictionaries (there are a number of edge cases: attributes, lists, anonymous lists, anonymous dicts, eval, etc. that most converters don't handle). In general, though, the converters are easy to use. If an 'example.xml' contains:

<top>
  <a>1</a>
  <b>2.2</b>
  <c>three</c>
</top>

Then to convert it to a dictionary:

>>> from xmlloader import *
>>> example = file('example.xml', 'r')   # A document containing XML
>>> xl = StreamXMLLoader(example, 0)     # 0 = all defaults on operation
>>> result = xl.expect XML()
>>> print result
{'top': {'a': '1', 'c': 'three', 'b': '2.2'}}

There are tools for converting in both C++ and Python: the C++ and Python do indentical conversion, but the C++ is about 60x faster

I have a recursive method to get a dictionary from a lxml element

    def recursive_dict(element):
        return (element.tag.split('}')[1],
                dict(map(recursive_dict, element.getchildren()),
                     **element.attrib))