Parse XML files in Python (ElementTree)

Question

I am trying to parse some xml and rss feeds to extract some of their data in order to store it in a PostGIS database.

The file I want to parse is here: http://earthquake.usgs.gov/earthquakes/catalogs/1hour-M1.xml and looks like this:

<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom" xmlns:georss="http://www.georss.org/georss">
  <updated>2012-03-05T19:57:55Z</updated>
  <title>USGS M 1+ Earthquakes</title>
  <subtitle>Real-time, worldwide earthquake list for the past hour</subtitle>
  <link rel="self" href="http://earthquake.usgs.gov/earthquakes/catalogs/1hour-M1.xml"/>
  <link href="http://earthquake.usgs.gov/earthquakes/"/>
  <author><name>U.S. Geological Survey</name></author>
  <id>http://earthquake.usgs.gov/</id>
  <icon>/favicon.ico</icon>
  <entry><id>urn:earthquake-usgs-gov:ak:10425830</id><title>M 2.5, Alaska Peninsula</title><updated>2012-03-05T19:37:00Z</updated><link rel="alternate" type="text/html" href="http://earthquake.usgs.gov/earthquakes/recenteqsww/Quakes/ak10425830.php"/><summary type="html"><![CDATA[<img src="http://earthquake.usgs.gov/images/globes/60_-155.jpg" alt="58.558&#176;N 155.673&#176;W" align="left" hspace="20" /><p>Monday, March  5, 2012 19:37:00 UTC<br>Monday, March  5, 2012 10:37:00 AM at epicenter</p><p><strong>Depth</strong>: 185.70 km (115.39 mi)</p>]]></summary><georss:point>58.5578 -155.6727</georss:point><georss:elev>-185700</georss:elev><category label="Age" term="Past hour"/></entry>
  <entry><id>urn:earthquake-usgs-gov:nc:71742560</id><title>M 1.5, San Francisco Bay area, California</title><updated>2012-03-05T19:31:16Z</updated><link rel="alternate" type="text/html" href="http://earthquake.usgs.gov/earthquakes/recenteqsus/Quakes/nc71742560.php"/><summary type="html"><![CDATA[<img src="http://earthquake.usgs.gov/images/globes/40_-120.jpg" alt="37.988&#176;N 122.455&#176;W" align="left" hspace="20" /><p>Monday, March  5, 2012 19:31:16 UTC<br>Monday, March  5, 2012 11:31:16 AM at epicenter</p><p><strong>Depth</strong>: 0.20 km (0.12 mi)</p>]]></summary><georss:point>37.9882 -122.4550</georss:point><georss:elev>-200</georss:elev><category label="Age" term="Past hour"/></entry>
  <entry><id>urn:earthquake-usgs-gov:ak:10425819</id><title>M 2.2, Central Alaska</title><updated>2012-03-05T19:28:00Z</updated><link rel="alternate" type="text/html" href="http://earthquake.usgs.gov/earthquakes/recenteqsus/Quakes/ak10425819.php"/><summary type="html"><![CDATA[<img src="http://earthquake.usgs.gov/images/globes/65_-150.jpg" alt="63.217&#176;N 150.524&#176;W" align="left" hspace="20" /><p>Monday, March  5, 2012 19:28:00 UTC<br>Monday, March  5, 2012 10:28:00 AM at epicenter</p><p><strong>Depth</strong>: 112.30 km (69.78 mi)</p>]]></summary><georss:point>63.2167 -150.5241</georss:point><georss:elev>-112300</georss:elev><category label="Age" term="Past hour"/></entry>
  <entry><id>urn:earthquake-usgs-gov:nc:71742550</id><title>M 1.8, Northern California</title><updated>2012-03-05T19:18:22Z</updated><link rel="alternate" type="text/html" href="http://earthquake.usgs.gov/earthquakes/recenteqsus/Quakes/nc71742550.php"/><summary type="html"><![CDATA[<img src="http://earthquake.usgs.gov/images/globes/40_-125.jpg" alt="38.818&#176;N 122.821&#176;W" align="left" hspace="20" /><p>Monday, March  5, 2012 19:18:22 UTC<br>Monday, March  5, 2012 11:18:22 AM at epicenter</p><p><strong>Depth</strong>: 2.40 km (1.49 mi)</p>]]></summary><georss:point>38.8177 -122.8205</georss:point><georss:elev>-2400</georss:elev><category label="Age" term="Past hour"/></entry>
  <entry><id>urn:earthquake-usgs-gov:ak:10425806</id><title>M 2.1, Southern Alaska</title><updated>2012-03-05T19:14:44Z</updated><link rel="alternate" type="text/html" href="http://earthquake.usgs.gov/earthquakes/recenteqsus/Quakes/ak10425806.php"/><summary type="html"><![CDATA[<img src="http://earthquake.usgs.gov/images/globes/60_-145.jpg" alt="60.501&#176;N 145.118&#176;W" align="left" hspace="20" /><p>Monday, March  5, 2012 19:14:44 UTC<br>Monday, March  5, 2012 10:14:44 AM at epicenter</p><p><strong>Depth</strong>: 17.10 km (10.63 mi)</p>]]></summary><georss:point>60.5011 -145.1175</georss:point><georss:elev>-17100</georss:elev><category label="Age" term="Past hour"/></entry>
</feed>

I am quite new to Python and to XML but I believe that the right direction to choose is ElementTree. Thus I have started with the following code:

#-*- coding: utf-8 -*-

import os
import urllib
import xml.etree.ElementTree as ET

def main():
  feed = urllib.urlopen("http://earthquake.usgs.gov/earthquakes/catalogs/1hour-M1.xml")

  try:
    tree = ET.parse(feed)
    print "Download ok"
    root = tree.getroot()
    print root
    event = root.find("entry")
    for e in event:
      print e.attrib
  except Exception, inst:
    print "Unexpected error opening %s: %s" % (tree, inst)

if __name__ == "__main__":
  main()

but an error is thrown...

Can anyone point me the best direction to follow for this? I do not really understand how to extract the data from the various tags. And then, what's the ideal strategy to "store" this data: an array of dictionaries?

Thanks in advance!

sgillies · Accepted Answer · 2012-03-06 18:55:14Z

Before I try to answer, a tip. Your exception handler covers up the nature of the problem. Just let the original exception rise up and you'll have more information to share with people who are interested in helping you.

I like to use feedparser to parse Atom feeds. It does indeed give you dict-like objects. I submitted a patch to feedparser 4.1 to parse the GeoRSS elements into GeoJSON style dicts. See https://code.google.com/p/feedparser/issues/detail?id=62 and blog post at http://sgillies.net/blog/566/georss-patch-for-universal-feedparser/. You'd use it like this:

>>> import feedparser
>>> feed = feedparser.parse("http://earthquake.usgs.gov/earthquakes/catalogs/1hour-M1.xml")
>>> feed.entries[0]['where']
{'type': 'Point', 'coordinates': (-122.8282, 38.844700000000003)}

My patched version of 4.1 is in my Dropbox and you can get it using pip.

$ pip install http://dl.dropbox.com/u/10325831/feedparser-4.1-georss.tar.gz

Or just download and install with "python setup.py install".

Thanks for your tip! I will also have a look at feedparser then! But can it parse any type of xml file? Or do you use it mainly / exclusively for atom feeds? — Stéphane Henriod, Mar 6 '12 at 20:09
Any type? No, but a surprising number. Atom, RSS 1 and 2, other feed related formats. — sgillies, Mar 6 '12 at 20:10

Alex Morega · Accepted Answer · 2012-03-06 17:04:00Z

It's more comfortable to use lxml for XML processing. Here is an example that fetches the feed and prints earthquake titles and coordinates:

import lxml.etree

feed_url = 'http://earthquake.usgs.gov/earthquakes/catalogs/1hour-M1.xml'
ns = {
    'atom': 'http://www.w3.org/2005/Atom',
    'georss': 'http://www.georss.org/georss',
}

def main():
    doc = lxml.etree.parse(feed_url)
    for entry in doc.xpath('//atom:entry', namespaces=ns):
        [title] = entry.xpath('./atom:title', namespaces=ns)
        [point] = entry.xpath('./georss:point', namespaces=ns)
        print point.text, title.text

if __name__ == '__main__':
    main()

lxml-etree is not recognized. I guess I have to install the package manually, right? ("ImportError: No module named lxml.etree") - Python 2.7 — Stéphane Henriod, Mar 6 '12 at 20:05
Yes, they have some installation instructions. If you're running under Linux your distribution may have lxml already packaged, e.g. Debian has the python-lxml package. — Alex Morega, Mar 6 '12 at 21:04

Brian · Accepted Answer · 2012-03-06 19:30:54Z

2

The problem you are having is a little tricky issue with ATOM feeds. When you parse the feed in Python, even though the xml simply shows the tag <entry> the actual tag is <{http://www.w3.org/2005/Atom}entry> as Python reads it.

So using the same code you posted above:

#-*- coding: utf-8 -*-

import os
import urllib
import xml.etree.ElementTree as ET

def main():
  feed = urllib.urlopen("http://earthquake.usgs.gov/earthquakes/catalogs/1hour-M1.xml")

  try:
    tree = ET.parse(feed)
    print "Download ok"
    root = tree.getroot()
    print root
    event = root.find("{http://www.w3.org/2005/Atom}entry")   ## Notice change here
    for e in event:
      print e.attrib
  except Exception, inst:
    print "Unexpected error opening %s: %s" % (tree, inst)

if __name__ == "__main__":
  main()

The error you are receiving is due to the fact that no entries are returned when you use event = root.find("entry") so when you try to loop through event Python returns a TypeError because you cannot loop through a variable that is equal to None.

Hope this helps.

answered Mar 6 '12 at 19:30

Brian

2,8491 gold badge25 silver badges47 bronze badges

This helps, thank you! No error is thrown, but I am still a bit lost with how to actually extract the data fields. What should I change to this code to get a dictionary, an array or something containing the values for each event (in this file, for each "entry")? – Stéphane Henriod Mar 6 '12 at 20:04
I would use root.findall instead of root.find. This will give you a list of all entries. Then you can loop through that list and add each entry to a dictionary. Use a nested loop through each entry to get the data. This should look like: for entry in root.findall("{http://www.w3.org/2005/Atom}entry"):\n adict[entry] = {}\n for e in entry.getchildren():\n adict[entry][e.tag] = e.text – Brian Mar 6 '12 at 20:26
Ok, this helps a lot, but I still have some trouble to loop through adict... Can you tell me what is wrong here? def main(): feed = urllib.urlopen("earthquake.usgs.gov/earthquakes/catalogs/1hour-M1.xml") tree = ET.parse(feed) root = tree.getroot() adict = {} for entry in root.findall("{w3.org/2005/Atom}entry"): adict[entry] = {} for e in entry.getchildren(): adict[entry] [e.tag] = e.text for item in adict: print item The program ends without any error and without printing anything... – Stéphane Henriod Mar 6 '12 at 21:03
PS: sorry for the poor formatting. Haven't understood how to cleanly format the comments... – Stéphane Henriod Mar 6 '12 at 21:07

add a comment |

David Bitner · Accepted Answer · 2012-03-06 16:39:11Z

1

Not an answer to be able to do this with Python, but to directly import data from a GeoRSS file you can use the ogr2ogr tool that is part of the OGR/GDAL package.

For further information on OGR drivers see:

GeoRSS Driver

PostGIS Driver

answered Mar 6 '12 at 16:39

David Bitner

5112 silver badges4 bronze badges

Thanks that's interesting, but I guess it only works with spatially enabled feeds? At the end, I will have to parse all sorts of xml / rss / atom feeds, not all of them having lat/long fields. I guess that Python is still the most versatile solution for that? – Stéphane Henriod Mar 6 '12 at 20:07

add a comment |

Stack Exchange Network

current community

your communities

more stack exchange communities

Parse XML files in Python (ElementTree)

4 Answers 4

Your Answer

Not the answer you're looking for? Browse other questions tagged python xml or ask your own question.

Hot Network Questions

Parse XML files in Python (ElementTree)

4 Answers 4

Your Answer

Sign up or log in

Post as a guest

Not the answer you're looking for? Browse other questions tagged python xml or ask your own question.

Related

Hot Network Questions