此解决方案通过在您提供的描述之间base.xml并test.xml基于您提供的描述进行简化的区别来工作。差异结果是结合原始树的第三棵XML树。输出是差异，文件之间不匹配的行以颜色编码突出显示。

我希望您可以使用它或使其适应您的需求。

复制粘贴脚本

import copy
from lxml import etree
import Tkinter as tk


# assumption: the root element of both trees is the same
# note: missing subtrees will only have the parent element highlighted


def element_content_equal(e1, e2):
    # starting point here: http://stackoverflow.com/a/24349916/377366
    try:
        if e1.tag != e1.tag:
            return False
        elif e1.text != e2.text:
            return False
        elif e1.tail != e2.tail:
            return False
        elif e1.attrib != e2.attrib:
            return False
    except AttributeError:
        # e.g. None is passed in for an element
        return False
    return True


def element_is_in_sequence(element, sequence):
    for e in sequence:
        if element_content_equal(e, element):
            return True
    return False


def copy_element_without_children(element):
    e_copy = etree.Element(element.tag, attrib=element.attrib, nsmap=element.nsmap)
    e_copy.text = element.text
    e_copy.tail = element.tail
    return e_copy


# start at the root of both xml trees
parser = etree.XMLParser(recover=True, remove_blank_text=True)
base_root = etree.parse('base.xml', parser=parser).getroot()
test_root = etree.parse('test.xml', parser=parser).getroot()
# each element from the original xml trees will be placed into a merge tree
merge_root = copy_element_without_children(base_root)


# additionally each merge tree element will be tagged with its source
DIFF_ATTRIB = 'diff'
FROM_BASE_ONLY = 'base'
FROM_TEST_ONLY = 'test'

# process the pair of trees, one set of parents at a time
parent_stack = [(base_root, test_root, merge_root)]
while parent_stack:
    base_parent, test_parent, merge_parent = parent_stack.pop()
    base_children = base_parent.getchildren()
    test_children = test_parent.getchildren()

    # compare children and transfer to merge tree
    base_children_iter = iter(base_children)
    test_children_iter = iter(test_children)
    base_child = next(base_children_iter, None)
    test_child = next(test_children_iter, None)
    while (base_child is not None) or (test_child is not None):
        # first handle the case of a unique base child
        if (base_child is not None) and (not element_is_in_sequence(base_child, test_children)):
            # base_child is unique: deep copy with base only tag
            merge_child = copy.deepcopy(base_child)
            merge_child.attrib[DIFF_ATTRIB] = FROM_BASE_ONLY
            merge_parent.append(merge_child)
            # this unique child has already been fully copied to the merge tree so it doesn't go on the stack
            # only move the base child since test child hasn't been handled yet
            base_child = next(base_children_iter, None)
        elif (test_child is not None) and (not element_is_in_sequence(test_child, base_children)):
            # test_child is unique: deep copy with base only tag
            merge_child = copy.deepcopy(test_child)
            merge_child.attrib[DIFF_ATTRIB] = FROM_TEST_ONLY
            merge_parent.append(merge_child)
            # this unique child has already been fully copied to the merge tree so it doesn't go on the stack
            # only move test child since base child hasn't been handled yet
            test_child = next(test_children_iter, None)
        elif element_content_equal(base_child, test_child):
            # both trees share the same element: shallow copy either child with shared tag
            merge_child = copy_element_without_children(base_child)
            merge_parent.append(merge_child)
            # put pair of children on stack as parents to be tested since their children may differ
            parent_stack.append((base_child, test_child, merge_child))
            # move on to next children in both trees since this was a shared element
            base_child = next(base_children_iter, None)
            test_child = next(test_children_iter, None)
        else:
            raise RuntimeError  # there is something wrong - element should be unique or shared.

# display merge_tree with highlighting to indicate source of each line
#   no highlight: common element in both trees
#   green: line that exists only in test tree (i.e. additional)
#   red: line that exists only in the base tree (i.e. missing)
root = tk.Tk()
textbox = tk.Text(root)
textbox.pack(expand=1, fill=tk.BOTH)
textbox.tag_config(FROM_BASE_ONLY, background='#ff5555')
textbox.tag_config(FROM_TEST_ONLY, background='#55ff55')

# find diff lines to highlight within merge_tree string that includes kludge attributes
merge_tree_string = etree.tostring(merge_root, pretty_print=True)
diffs_by_line = []
for line, line_text in enumerate(merge_tree_string.split('\n')):
    for diff_type in (FROM_BASE_ONLY, FROM_TEST_ONLY):
        if diff_type in line_text:
            diffs_by_line.append((line+1, diff_type))

# remove kludge attributes
for element in merge_root.iter():
    try:
        del(element.attrib[DIFF_ATTRIB])
    except KeyError:
        pass
merge_tree_string = etree.tostring(merge_root, pretty_print=True)

# highlight final lines
textbox.insert(tk.END, merge_tree_string)
for line, diff_type in diffs_by_line:
    textbox.tag_add(diff_type, '{}.0'.format(line), '{}.0'.format(int(line)+1))
root.mainloop()

输入：

请注意，我清理了xml，因为与原始XML的行为不一致。原始的基本上是使用反斜杠代替正斜杠，并且在打开标签上也使用了虚假的结束斜杠。

base.xml （与此脚本位于同一位置）

<?xml version="1.0" encoding="UTF-8" standalone="no" ?>
<p1:sample1 xmlns:p1="http://www.example.org/eHorizon">
   <p1:time nTimestamp="5">
      <p1:location hours = "1" path = '1'>
         <p1:feature color="6" type="a">560</p1:feature>
         <p1:feature color="2" type="a">564</p1:feature>
         <p1:feature color="3" type="b">570</p1:feature>
         <p1:feature color="4" type="c">570</p1:feature>
      </p1:location>
      <p1:location hours = "5" path = '1'>
         <p1:feature color="6" type="a">560</p1:feature>
         <p1:feature color="7" type="b">570</p1:feature>
         <p1:feature color="8" type="c">580</p1:feature>
      </p1:location>
      <p1:location hours = "5" path = '1'>
         <p1:feature color="6" type="a">560</p1:feature>
      </p1:location>
   </p1:time>
   <p1:time nTimestamp="6">
      <p1:location hours = "1" path = '1'>
         <p1:feature color="2" type="a">564</p1:feature>
         <p1:feature color="3" type="b">570</p1:feature>
         <p1:feature color="4" type="c">570</p1:feature>
      </p1:location>
      <p1:location hours = "5" path = '1'>
         <p1:feature color="6" type="a">560</p1:feature>
         <p1:feature color="9" type="b">590</p1:feature>
         <p1:feature color="10" type="c">600</p1:feature>
      </p1:location>
      <p1:location hours = "5" path = '1'>
         <p1:feature color="6" type="a">560</p1:feature>
         <p1:feature color="7" type="b">570</p1:feature>
         <p1:feature color="8" type="c">580</p1:feature>
      </p1:location>
   </p1:time>
</p1:sample1>

test.xml （与此脚本位于同一位置）

<?xml version="1.0" encoding="UTF-8" standalone="no" ?>
<p1:sample1 xmlns:p1="http://www.example.org/eHorizon">
   <p1:time nTimestamp="5">
      <p1:location hours = "1" path = '1'>
         <p1:feature color="6" type="a">560</p1:feature>
         <p1:feature color="2" type="a">564</p1:feature>
         <p1:feature color="3" type="b">570</p1:feature>
         <p1:feature color="4" type="c">570</p1:feature>
      </p1:location>
      <p1:location hours = "5" path = '1'>
         <p1:feature color="6" type="a">560</p1:feature>
         <p1:feature color="7" type="b">570</p1:feature>
         <p1:feature color="8" type="c">580</p1:feature>
      </p1:location>
      <p1:location hours = "5" path = '1'>
         <p1:feature color="9" type="b">1111</p1:feature>
         <p1:feature color="10" type="c">2222</p1:feature>
      </p1:location>
   </p1:time>
   <p1:time nTimestamp="6">
      <p1:location hours = "1" path = '1'>
         <p1:feature color="2" type="a">564</p1:feature>
         <p1:feature color="3" type="b">570</p1:feature>
         <p1:feature color="4" type="c">570</p1:feature>
      </p1:location>
      <p1:location hours = "5" path = '1'>
         <p1:feature color="6" type="a">560</p1:feature>
         <p1:feature color="9" type="b">590</p1:feature>
         <p1:feature color="10" type="c">600</p1:feature>
      </p1:location>
      <p1:location hours = "5" path = '1'>
         <p1:feature color="6" type="a">560</p1:feature>
         <p1:feature color="7" type="b">570</p1:feature>
         <p1:feature color="8" type="c">580</p1:feature>
      </p1:location>
   </p1:time>
</p1:sample1>