Python Manipulating XML

A short script with descriptions for reading and manipulating xml. It seems like the python ElementTree module should be the easiest and best suited for XML manipulation. However I had a complicated XML structure with multiple namespaces and lxml handled it better. ElementTree could only handle one namespace with it;s register function. In addition lxml has pretty_print which might be useful. Although in my case when I do inserts pretty_print did not work even with the FAX fix for remove_blank_text.

1
import lxml.etree as ET
2

3
f = open('config.xml','rb')
4
## http://lxml.de/FAQ.html#why-doesn-t-the-pretty-print-option-reformat-my-xml-output
5
#parser = ET.XMLParser(remove_blank_text=True)
6
#tree = ET.parse(f, parser)
7
tree = ET.parse(f)
8

9
#for element in tree.iter():
10
#    element.tail = None
11

12
root = tree.getroot()
13
namespace="http://xmlns.oracle.com/weblogic/domain"
14
servers = tree.findall('.//{%s}server' % namespace)
15

16
## Loop through the nodes we found
17
for server in servers:
18
  print "New SERVER node detected:"
19
  for child in server:
20
    tag = child.tag
21
    val = child.text
22
    ## Remove any existing children
23
    if tag == "{http://xmlns.oracle.com/weblogic/domain}ssl":
24
      print "found server.ssl and will remove",
25
      server.remove(child)
26
    if tag == "{http://xmlns.oracle.com/weblogic/domain}log":
27
      print "found server.log and will remove",
28
      server.remove(child)
29
    if tag == "{http://xmlns.oracle.com/weblogic/domain}data-source":
30
      print "found server.data-source and will remove",
31
      server.remove(child)
32
    print tag, val
33

34
  ## Add the 3 children we want
35
  child = ET.Element("ssl")
36
  child.text=''
37
  server.insert(1,child)
38
  ##  Check out why xsi:nil is not working. UTF???
39
  ##  gchild = ET.Element("hostname-verifier",attrib={'xsi:nil':'true'})
40
  gchild = ET.Element("hostname-verifier",attrib={'xsi_nil':'true'})
41
  gchild.text=''
42
  child.insert(1,gchild)
43
  gchild = ET.Element("hostname-verification-ignored")
44
  gchild.text='true'
45
  child.insert(2,gchild)
46
  gchild = ET.Element("client-certificate-enforced")
47
  gchild.text='true'
48
  child.insert(3,gchild)
49
  gchild = ET.Element("two-way-ssl-enabled")
50
  gchild.text='false'
51
  child.insert(3,gchild)
52

53
  child = ET.Element("log")
54
  child.text=''
55
  server.insert(2,child)
56
  gchild = ET.Element("rotation-type")
57
  gchild.text='byTime'
58
  child.insert(1,gchild)
59
  gchild = ET.Element("number-of-files-limited")
60
  gchild.text='true'
61
  child.insert(2,gchild)
62
  gchild = ET.Element("rotate-log-on-startup")
63
  gchild.text='true'
64
  child.insert(3,gchild)
65

66
  child = ET.Element("data-source")
67
  child.text=''
68
  server.insert(3,child)
69
  gchild = ET.Element("data-source-log-file")
70
  gchild.text=''
71
  child.insert(1,gchild)
72
  ggchild = ET.Element("rotation-type")
73
  ggchild.text='byTime'
74
  gchild.insert(1,ggchild)
75
  ggchild = ET.Element("number-of-files-limited")
76
  ggchild.text='true'
77
  gchild.insert(2,ggchild)
78
  ggchild = ET.Element("rotate-log-on-startup")
79
  ggchild.text='true'
80
  gchild.insert(3,ggchild)
81

82
## Check out why pretty_print is not making newlines in new tags
83
#print(ET.tostring(tree, pretty_print=True))
84
tree.write("wc-out.xml", pretty_print=True)