Should I use PyXML or what's in the standard library?
Add a comment
|
3 Answers
ElementTree is provided as part of the standard Python libs. ElementTree is pure python, and cElementTree is the faster C implementation:
# Try to use the C implementation first, falling back to python
try:
from xml.etree import cElementTree as ElementTree
except ImportError, e:
from xml.etree import ElementTree
Here's an example usage, where I'm consuming xml from a RESTful web service:
def find(*args, **kwargs):
"""Find a book in the collection specified"""
search_args = [('access_key', api_key),]
if not is_valid_collection(kwargs['collection']):
return None
kwargs.pop('collection')
for key in kwargs:
# Only the first keword is honored
if kwargs[key]:
search_args.append(('index1', key))
search_args.append(('value1', kwargs[key]))
break
url = urllib.basejoin(api_url, '%s.xml' % 'books')
data = urllib.urlencode(search_args)
req = urllib2.urlopen(url, data)
rdata = []
chunk = 'xx'
while chunk:
chunk = req.read()
if chunk:
rdata.append(chunk)
tree = ElementTree.fromstring(''.join(rdata))
results = []
for i, elem in enumerate(tree.getiterator('BookData')):
results.append(
{'isbn': elem.get('isbn'),
'isbn13': elem.get('isbn13'),
'title': elem.find('Title').text,
'author': elem.find('AuthorsText').text,
'publisher': elem.find('PublisherText').text,}
)
return results
3 Comments
rick
vezult, how come sometimes you use elem.get() and sometimes you use elem.find().text?
vezult
@rick: elem.get() is fetching the value of an element attribute, while elem.find() is searching for elements contained within the elem element.
jfs
tree = ElementTree.parse(urllib2.urlopen(url, data)) should work without rdata list.I always prefer to use the standard library when possible. ElementTree is well known amongst pythonistas, so you should be able to find plenty of examples. Parts of it have also been optimized in C, so it's quite fast.
Comments
There's also BeautifulSoup, which has an API some might prefer. Here's an example on how you can extract all tweets that have been favorited from Twitter's Public Timeline:
from BeautifulSoup import BeautifulStoneSoup
import urllib
url = urllib.urlopen('http://twitter.com/statuses/public_timeline.xml').read()
favorited = []
soup = BeautifulStoneSoup(url)
statuses = soup.findAll('status')
for status in statuses:
if status.find('favorited').contents != [u'false']:
favorited.append(status)
3 Comments
mlissner
Alas, BeautifulSoup is no longer maintained. I would avoid it, and lean towards lxml or ElementTree.
mrkzq
@mlissner I can't see where does it says on BS4 website that it is no longer maintained. Is that really the case?
mlissner
At one point the maintainer was threatening to step down, but it seems that reality never came to pass.