Skip to content

Instantly share code, notes, and snippets.

@dalelane
Created October 6, 2014 02:03
Show Gist options
  • Save dalelane/a0514b2e283a882d9ef3 to your computer and use it in GitHub Desktop.
Save dalelane/a0514b2e283a882d9ef3 to your computer and use it in GitHub Desktop.
Comparing XML files ignoring order of attributes and elements - see http://dalelane.co.uk/blog/?p=3225 for background
##########################################################################
#
# xmldiff
#
# Simple utility script to enable a diff of two XML files in a way
# that ignores the order or attributes and elements.
#
# Dale Lane (email@dalelane.co.uk)
# 6 Oct 2014
#
##########################################################################
#
# Overview
# The approach is to sort both files by attribute and element, and
# then reuse an existing diff implementation on the sorted files.
#
# Arguments
# <diffcommand> the command that should be run to diff the sorted files
# <filename1> the first XML file to diff
# <filename2> the second XML file to diff
#
# Background
# http://dalelane.co.uk/blog/?p=3225
#
##########################################################################
import os, sys, subprocess, platform
import lxml.etree as le
from operator import attrgetter
#
# Check required arguments
if len(sys.argv) != 4:
print ("Usage: python xmldiff.py <diffcommand> <filename1> <filename2>")
quit()
#
# Prepares the location of the temporary file that will be created by xmldiff
def createFileObj(prefix, name):
return {
"filename" : os.path.abspath(name),
"tmpfilename" : "." + prefix + "." + os.path.basename(name)
}
#
# Function to sort XML elements by id
# (where the elements have an 'id' attribute that can be cast to an int)
def sortbyid(elem):
id = elem.get('id')
if id:
try:
return int(id)
except ValueError:
return 0
return 0
#
# Function to sort XML elements by their text contents
def sortbytext(elem):
text = elem.text
if text:
return text
else:
return ''
#
# Function to sort XML attributes alphabetically by key
# The original item is left unmodified, and it's attributes are
# copied to the provided sorteditem
def sortAttrs(item, sorteditem):
attrkeys = sorted(item.keys())
for key in attrkeys:
sorteditem.set(key, item.get(key))
#
# Function to sort XML elements
# The sorted elements will be added as children of the provided newroot
# This is a recursive function, and will be called on each of the children
# of items.
def sortElements(items, newroot):
# The intended sort order is to sort by XML element name
# If more than one element has the same name, we want to
# sort by their text contents.
# If more than one element has the same name and they do
# not contain any text contents, we want to sort by the
# value of their ID attribute.
# If more than one element has the same name, but has
# no text contents or ID attribute, their order is left
# unmodified.
#
# We do this by performing three sorts in the reverse order
items = sorted(items, key=sortbyid)
items = sorted(items, key=sortbytext)
items = sorted(items, key=attrgetter('tag'))
# Once sorted, we sort each of the items
for item in items:
# Create a new item to represent the sorted version
# of the next item, and copy the tag name and contents
newitem = le.Element(item.tag)
if item.text and item.text.isspace() == False:
newitem.text = item.text
# Copy the attributes (sorted by key) to the new item
sortAttrs(item, newitem)
# Copy the children of item (sorted) to the new item
sortElements(list(item), newitem)
# Append this sorted item to the sorted root
newroot.append(newitem)
#
# Function to sort the provided XML file
# fileobj.filename will be left untouched
# A new sorted copy of it will be created at fileobj.tmpfilename
def sortFile(fileobj):
with open(fileobj['filename'], 'r') as original:
# parse the XML file and get a pointer to the top
xmldoc = le.parse(original)
xmlroot = xmldoc.getroot()
# create a new XML element that will be the top of
# the sorted copy of the XML file
newxmlroot = le.Element(xmlroot.tag)
# create the sorted copy of the XML file
sortAttrs(xmlroot, newxmlroot)
sortElements(list(xmlroot), newxmlroot)
# write the sorted XML file to the temp file
newtree = le.ElementTree(newxmlroot)
with open(fileobj['tmpfilename'], 'wb') as newfile:
newtree.write(newfile, pretty_print=True)
#
# sort each of the specified files
filefrom = createFileObj("from", sys.argv[2])
sortFile(filefrom)
fileto = createFileObj("to", sys.argv[3])
sortFile(fileto)
#
# invoke the requested diff command to compare the two sorted files
if platform.system() == "Windows":
sp = subprocess.Popen([ "cmd", "/c", sys.argv[1] + " " + filefrom['tmpfilename'] + " " + fileto['tmpfilename'] ])
sp.communicate()
else:
sp = subprocess.Popen([ "/bin/bash", "-i", "-c", sys.argv[1] + " " + os.path.abspath(filefrom['tmpfilename']) + " " + os.path.abspath(fileto['tmpfilename']) ])
sp.communicate()
#
# cleanup - delete the temporary sorted files after the diff terminates
os.remove(filefrom['tmpfilename'])
os.remove(fileto['tmpfilename'])
@ashish1405
Copy link

this works nicely just for osx the popen command should be as below
sp = subprocess.Popen([ "open", "-a", sys.argv[1] , "--args" , os.path.abspath(filefrom['tmpfilename']) , os.path.abspath(fileto['tmpfilename']) ])

@achyan
Copy link

achyan commented Jun 30, 2016

This breaks if there are any comments. I added in sortElements:

    # Once sorted, we sort each of the items
    for item in items:

        **if item.tag is le.Comment:
            continue**

@davidbd
Copy link

davidbd commented Aug 26, 2016

I am relatively new to using the terminal on the Mac. When I try your command line instruction python xmldiff.py diffmerge FileA.xml FileB.xml it returns the error "bash: diffmerge: command not found". What am I doing wrong?

@allanlewis
Copy link

I've made this into a repo here: https://github.com/allanlewis/xmldiff

@Mhussein27
Copy link

Is there any Git Client that can support this function?
I need Xml Ignore order function to work when I am comparing my XML through tortoisegit or any client.

@pombredanne
Copy link

@dalelane what would be the license for this code?

@anoop1984
Copy link

it is not working for following set of files
1.xml


2
rnc1
na0
1
ssn
1
0x21a7


1
rnc1
na0
1
ssn
1
0x21a7


3
rnc1
na0
1
ssn
1
0x21a7


2.xml


2
rnc1
na0
1
ssn
1
0x21a7


3
rnc1
na0
1
ssn
1
0x21a7


1
rnc1
na0
1
ssn
1
0x21a7


even the files are same just the input order for load-sharing-item is changed, but still it showing the difference
$ python Diff_xml.py diff 1.xml 2.xml
14c14
< 1

<load-sharing-id>3</load-sharing-id>

23c23
< 3

<load-sharing-id>1</load-sharing-id>

@romaad
Copy link

romaad commented Dec 25, 2017

following error happens to me:
lxml.etree.XMLSyntaxError: Start tag expected, '<' not found, line 1, column 1
from line 126
even though the file starts with a '<'

@abeaclark
Copy link

Great code, thanks! Would you mind adding a license to it? I would love to use this, but don't want to if it's not licensed for re-use!

@matkoniecz
Copy link

@dalelane what would be the license for this code? See http://choosealicense.com/ for help.

@HarryF514
Copy link

I am getting this error:

Traceback (most recent call last):
  File "xmldiff.py", line 146, in <module>
    sortFile(filefrom)
  File "xmldiff.py", line 135, in sortFile
    sortElements(list(xmlroot), newxmlroot)
  File "xmldiff.py", line 105, in sortElements
    newitem = le.Element(item.tag)
  File "src/lxml/etree.pyx", line 2996, in lxml.etree.Element
  File "src/lxml/apihelpers.pxi", line 95, in lxml.etree._makeElement
  File "src/lxml/apihelpers.pxi", line 1584, in lxml.etree._getNsTag
  File "src/lxml/apihelpers.pxi", line 1602, in lxml.etree.__getNsTag
  File "src/lxml/apihelpers.pxi", line 1472, in lxml.etree._utf8
TypeError: Argument must be bytes or unicode, got 'cython_function_or_method'

@Gnafu
Copy link

Gnafu commented Feb 28, 2019

@harryyuanfeng you may have comments in your xml.
You can ignore them by changing the "sortfile" function:

        parser = le.XMLParser(remove_comments=True)
        # parse the XML file and get a pointer to the top
        xmldoc = le.parse(original, parser=parser)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment