Question

My question revolves around a XML file I am working with it looks like this.

<log>
<logentry
 revision="33185">
<author>glv</author>
<date>2012-08-06T21:01:52.494219Z</date>
<paths>

<path
 kind="file"
 action="M">/branches/Patch_4_2_0_Branch/text.xml</path>   

<path
 kind="dir"
 action="M">/branches/Patch_4_2_0_Branch</path>

</paths>
<msg>PATCH_BRANCH:N/A
BUG_NUMBER:N/A
FEATURE_AFFECTED:N/A
OVERVIEW:N/A
Adding the SVN log size requirement to the branch 
</msg>
  </logentry>
    </log>

Now what I want to do is I want to use a “if” statement on this to look in the xml path tag to check if it is a kind= dir or a kind = file. and then add the path to a varible called content. This is what I have so far for it. I am using dom.import btw.

xmlPath = dom.getElementsByTagName('paths')[0]
xmlPathM =  xmlPath.getAttribute('kind')    
if xmlPathM == dir:
    content += "Directory location:" + xmlPathM +"\n \n"
else:
    content += "FileName"  + xmlPathM +"\n \n "

Now it doesn’t seem to want to work it will print out FileName in it but not the Directory location. I believe but i want it to look at this logentry and print out this

Directory location: /branches/Patch_4_2_0_Branch 

FileName:/branches/Patch_4_2_0_Branch/text.xml

for the same log-entry.
Any idea as to what I am missing or doing wrong?

Answer 1

Working with XML is a great thing to do with a good interactive console, like IPython. [Side note: I prefer ElementTree because I like the interface more, but whatever]

Let’s dive in. First read it in:

In [51]: import xml.dom.minidom as minidom

In [52]: dom = minidom.parse("log.xml")

In [53]: dom
Out[53]: <xml.dom.minidom.Document instance at 0x97082ec>

Now look inside the paths:

In [55]: dom.getElementsByTagName("paths")
Out[55]: [<DOM Element: paths at 0x97086cc>]

In [56]: dom.getElementsByTagName("paths")[0]
Out[56]: <DOM Element: paths at 0x97086cc>

In [57]: vars(dom.getElementsByTagName("paths")[0])
Out[57]: 
{'_attrs': {},
 '_attrsNS': {},
 'childNodes': [<DOM Text node "u'\n\n'">,
  <DOM Element: path at 0x970874c>,
  <DOM Text node "u'   \n\n'">,
  <DOM Element: path at 0x97088ac>,
  <DOM Text node "u'\n\n'">],
 'namespaceURI': None,
 'nextSibling': <DOM Text node "u'\n'">,
 'nodeName': u'paths',
 'ownerDocument': <xml.dom.minidom.Document instance at 0x97082ec>,
 'parentNode': <DOM Element: logentry at 0x970848c>,
 'prefix': None,
 'previousSibling': <DOM Text node "u'\n'">,
 'tagName': u'paths'}

Look at the childNodes:

In [58]: dom.getElementsByTagName("paths")[0].childNodes
Out[58]: 
[<DOM Text node "u'\n\n'">,
 <DOM Element: path at 0x970874c>,
 <DOM Text node "u'   \n\n'">,
 <DOM Element: path at 0x97088ac>,
 <DOM Text node "u'\n\n'">]

Whitespace is significant, so that’s a bit of a headache. Can throw the non-elements away, though:

In [61]: elements = [x for x in dom.getElementsByTagName("paths")[0].childNodes if isinstance(x, minidom.Element)]

In [62]: elements
Out[62]: [<DOM Element: path at 0x970874c>, <DOM Element: path at 0x97088ac>]

Looking inside:

In [65]: elements
Out[65]: [<DOM Element: path at 0x970874c>, <DOM Element: path at 0x97088ac>]

In [66]: vars(elements[0])
Out[66]: 
{'_attrs': {u'action': <xml.dom.minidom.Attr instance at 0x970880c>,
  u'kind': <xml.dom.minidom.Attr instance at 0x97087ac>},
 '_attrsNS': {(None, u'action'): <xml.dom.minidom.Attr instance at 0x970880c>,
  (None, u'kind'): <xml.dom.minidom.Attr instance at 0x97087ac>},
 'childNodes': [<DOM Text node "u'/branches/'...">],
 'namespaceURI': None,
 'nextSibling': <DOM Text node "u'   \n\n'">,
 'nodeName': u'path',
 'ownerDocument': <xml.dom.minidom.Document instance at 0x97082ec>,
 'parentNode': <DOM Element: paths at 0x97086cc>,
 'prefix': None,
 'previousSibling': <DOM Text node "u'\n\n'">,
 'tagName': u'path'}

And finally we know what we want:

In [67]: for elem in elements:
    print elem, elem.childNodes[0].nodeValue, elem.getAttribute("kind"), elem.getAttribute("action") 
   ....:     
<DOM Element: path at 0x970874c> /branches/Patch_4_2_0_Branch/text.xml file M
<DOM Element: path at 0x97088ac> /branches/Patch_4_2_0_Branch dir M

I can’t imagine not doing this interactively.