scrapy response.xpath使用默认命名空间返回xml文档上的空数组，而response.re工作

Question

I am new to scrappy and I was playing with the scrapy shell trying to crawl this site: www.spiegel.de/sitemap.xml

I did it with

scrapy shell "http://www.spiegel.de/sitemap.xml"

and it works all fine, when i use

response.body 

i can see the whole page including xml tags

however for instance this:

response.xpath('//loc') 

simply wont work.

The result i get is an empty array

while

response.selector.re('somevalidregexpexpression') 

would work

any idea what could be the reason? could be related to encoding or so? the site is not utf-8

I am using python 2.7 on Win 7. I tried the xpath() on another site (dmoz) and it worked fine.

Answer 1

The problem was due to the default namespace declared at the root element of the XML :

xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"

So in that XML, the root element and its descendants without prefix inherits the same namespace, implicitly .

On the other hand, in XPath, you need to use prefix that bound to a namespace URI to reference element in that namespace, there is no such default namespace implied.

You can use selector.register_namespace() to bind a namespace prefix to the default namespace URI, and then use the prefix in your XPath :

response.selector.register_namespace('d', 'http://www.sitemaps.org/schemas/sitemap/0.9')
response.xpath('//d:loc')

Answer 2

You can also use xpath with local namespace such as in:

response.xpath("//*[local-name()='loc']")

This is especially useful if you are parsing responses from multiple heterogeneous sources and you don't want to register each and every namespace.