Python lxml: Strange issue with reading repeated elements and storing in a list

I'm running into a bizarre issue. I've had two very different versions of code to solve the same issue, and have run into the same problem.

I have simplified the problem down to this:

Here is the xml file:

<Test>
 <Object name="Ob1">
  <List/>
 </Object>

 <Object name="Ob2">
  <List>
   <item>One</item>
   <item>Two</item>
  </List>
 </Object>

 <Object name="Ob3">
  <List>
   <item>Three</item开发者_运维知识库>
   <item>Four</item>
   <item>Five</item>
  </List>
 </Object>
</Test>

Here is the python code:

from lxml import etree

#Load XML

fileobject = open("list_test.xml", "r") #read-only
parser = etree.XMLParser(remove_blank_text=True)
tree = etree.parse(fileobject, parser)
root = tree.getroot()

object_list = []

class TestClass():
    name = None
    list = []


for OB in root:
    temp_ob = TestClass()
    temp_ob.name = OB.get("name")
    for SubElem in OB:
        if SubElem.tag == "List":
            for item in SubElem:
                if item.tag == "item":
                    temp_ob.list.append(item.text)
    object_list.append(temp_ob)
    del temp_ob

for ob in object_list:
    print ob.name
    print ob.list

The code is supposed to store all the <item> elements in a list in an object, which is itself stored in a list.

However, here is the output I am getting:

Ob1
['One', 'Two', 'Three', 'Four', 'Five']
Ob2
['One', 'Two', 'Three', 'Four', 'Five']
Ob3
['One', 'Two', 'Three', 'Four', 'Five']

Why is it getting all the <item> elements in the WHOLE document?

The TestClass.list is a class-level attribute, so each object_list.append() is happening on the same list.

For example:

class Foo(object):
    lst = []

f1 = Foo()
f1.lst.append(1)
f2 = Foo()
f2.lst.append(2)
print f1.lst
print f2.lst

[1, 2]
[1, 2]

You should make it an instance-level attribute:

class Bar(object):
    def __init__(self):
        self.lst = []

b1 = Bar()
b1.lst.append(1)
b2 = Bar()
b2.lst.append(2)
print b1.lst
print b2.lst

[1]
[2]