How to sort XML by attribute (strings before numbers) in Python

Question

I would like to sort the below xml, by the attribute "value" of the "entry" tags and sort the strings (letters) before the numbers.

<test>
    <entry value="-12" />
    <entry value="0" />
    <entry value="043" />
    <entry value="14" />
    <entry value="6" />
    <entry value="_null" />
    <entry value="abc" />
    <entry value="abcd" />
    <entry value="empty" />
    <entry value="false" />
    <entry value="test1" />
    <entry value="test2" />
    <entry value="true" />
</test>

I have written some python that sorts this xml, but it sorts first the numbers and then the strings. I have checked this thread, but could not implement any of the solutions to sorting XML.

import xml.etree.ElementTree as ElT
import os
from os.path import sep

def sort_xml(directory, xml_file, level1_tag, attribute, mode=0):
    #mode 0 - numbers before letters
    #mode 1 - letters before numbers

    file = directory + sep + xml_file

    tree = ElT.parse(file)
    data = tree.getroot()
    els = data.findall(level1_tag)
    
    if mode == 0:
        new_els = sorted(els, key=lambda e: (e.tag, e.attrib[attribute]))
    if mode == 1:
        new_els = sorted(els, key=lambda e: (isinstance(e.tag, (float, int)), e.attrib[attribute]))

    for el in new_els:
        if mode == 0:
            el[:] = sorted(el, key=lambda e: (e.tag, e.attrib[attribute]))
        if mode == 1:
            el[:] = sorted(el, key=lambda e: (isinstance(e.tag, (float, int)), e.attrib[attribute]))
    
    data[:] = new_els

    tree.write(file, xml_declaration=True, encoding='utf-8')

    with open(file, 'r') as fin:
        data = fin.read().splitlines(True)
    with open(file, 'w') as fout:
        fout.writelines(data[1:])
        
        
sort_xml(os.getcwd(), "test.xml", "entry", "value", 1)

Any ideas how this could be done?

Edit1: Desired output

<test>
    <entry value="_null" />
    <entry value="abc" />
    <entry value="abcd" />
    <entry value="empty" />
    <entry value="false" />
    <entry value="test1" />
    <entry value="test2" />
    <entry value="true" />
    <entry value="-12" />
    <entry value="0" />
    <entry value="043" />
    <entry value="14" />
    <entry value="6" />
</test>

_null is a string - why it goes after the numbers? why -12 is before /this? — balderman
– balderman, Commented Mar 11, 2021 at 10:54
@balderman. The way the xml is sorted right now: <entry value="-12" /> <entry value="/this" /> <entry value="0" /> <entry value="043" /> <entry value="14" /> <entry value="6" /> <entry value="_null" /> <entry value="abc" /> <entry value="abcd" /> <entry value="empty" /> <entry value="false" /> <entry value="test1" /> <entry value="test2" /> <entry value="true" /> I took the part where the letters start and put it at the top. This the actual requirement, to have the letters at the top, I don't care about the rest. — Mihail-Cosmin Munteanu
– Mihail-Cosmin Munteanu, Commented Mar 11, 2021 at 10:57

tomjn · Accepted Answer · 2021-03-11 11:31:23Z

I think your problem is that when you are sorting you are checking if the value is an int or float. In fact all the values are strings e.g. isinstance(e.tag, (float, int)) will always be false.

A sorter function like this does what you want

def sorter(x):
    "Check if the value can be interpreted as an integer, then by the string"
    value = x.get("value") 
    def is_integer(i):
        try:
            int(i)
        except ValueError:
            return False
        return True
    return is_integer(value), value

which can be used like so (using StringIO as a substitute for the file)

from xml.etree import ElementTree
from io import StringIO

xml = """<test>
    <entry value="-12" />
    <entry value="0" />
    <entry value="043" />
    <entry value="14" />
    <entry value="6" />
    <entry value="_null" />
    <entry value="abc" />
    <entry value="abcd" />
    <entry value="empty" />
    <entry value="false" />
    <entry value="test1" />
    <entry value="test2" />
    <entry value="true" />
</test>"""

tree = ElementTree.parse(StringIO(xml))
root = tree.getroot()
root[:] = sorted(root, key=sorter)
tree.write("output.xml")

The contents of output.xml is

<test>
    <entry value="_null" />
    <entry value="abc" />
    <entry value="abcd" />
    <entry value="empty" />
    <entry value="false" />
    <entry value="test1" />
    <entry value="test2" />
    <entry value="true" />
    <entry value="-12" />
    <entry value="0" />
    <entry value="043" />
    <entry value="14" />
    <entry value="6" />
</test>

Thanks, this works, and keeps all values exactly as they are. Doesn't change 043 to 43.

balderman · Accepted Answer · 2021-03-11 12:14:12Z

1

I took the part where the letters start and put it at the top. This the actual requirement to have the letters at the top, I don't care about the rest.

below

 import xml.etree.ElementTree as ET

xml = '''<test>
    <entry value="-12" />
    <entry value="/this" />
    <entry value="0" />
    <entry value="043" />
    <entry value="14" />
    <entry value="6" />
    <entry value="_null" />
    <entry value="abc" />
    <entry value="abcd" />
    <entry value="empty" />
    <entry value="false" />
    <entry value="test1" />
    <entry value="test2" />
    <entry value="true" />
</test>'''

root = ET.fromstring(xml)
numeric = []
non_numeric = []
for entry in root.findall('.//entry'):
    try:
        x = int(entry.attrib['value'])
        numeric.append((x, entry.attrib['value']))
    except ValueError as e:
        non_numeric.append(entry.attrib['value'])

sorted(numeric, key=lambda x: x[0])
sorted(non_numeric)

root = ET.Element('test')
for value in non_numeric:
    entry = ET.SubElement(root, 'entry')
    entry.attrib['value'] = value
for value in numeric:
    entry = ET.SubElement(root, 'entry')
    entry.attrib['value'] = str(value[1])
ET.dump(root)

output

 <?xml version="1.0" encoding="UTF-8"?>
<test>
   <entry value="/this" />
   <entry value="_null" />
   <entry value="abc" />
   <entry value="abcd" />
   <entry value="empty" />
   <entry value="false" />
   <entry value="test1" />
   <entry value="test2" />
   <entry value="true" />
   <entry value="-12" />
   <entry value="0" />
   <entry value="043" />
   <entry value="14" />
   <entry value="6" />
</test>

edited Mar 11, 2021 at 12:14

answered Mar 11, 2021 at 11:05

balderman

24k8 gold badges39 silver badges60 bronze badges

4 Comments

Mihail-Cosmin Munteanu Over a year ago

This works, but how can I save the output to a file?

balderman Over a year ago

see stackoverflow.com/questions/25856774/…

Mihail-Cosmin Munteanu Over a year ago

This solution changes some of the values, for example 043 becomes 43.

balderman Over a year ago

@Mihail-CosminMunteanu fixed. see latest code

Collectives™ on Stack Overflow

How to sort XML by attribute (strings before numbers) in Python

2 Answers 2

1 Comment

4 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

4 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related