使用 lxml 提取 xml 标记值，使用 Python 提取 xpath

Question

I have a requirement where I have to extract XML data using lxml and xpath. I need to extract EventId = 122157660 using lxml and xpath.

<B2B_DATA>
   <B2B_METADATA>
       <EventId>122157660</EventId>
       <MessageType>Request</MessageType>
   </B2B_METADATA>
<PAYLOAD>
    <![CDATA[<?xml version="1.0"?>
        <REQUEST_GROUP MISMOVersionID="1.1.1">
            <REQUESTING_PARTY _Name="CityBank" _StreetAddress="801 Main St" _City="rockwall" _State="MD" _PostalCode="11311" _Identifier="416">
                <CONTACT_DETAIL _Name="XX Davis">
                    <CONTACT_POINT _Type="Phone" _Value="1236573348"/>
                    <CONTACT_POINT _Type="Email" _Value="jXX@city.com"/>
                </CONTACT_DETAIL>
            </REQUESTING_PARTY>
        </REQUEST_GROUP>]]>
</PAYLOAD>
</B2B_DATA>

I am able to do this using loops and iter but would like use xpath for cleaner/shorter code. Also I am using lxml using to parse CDATA, so trying to avoid ElementTree lib.

This is what I tried -

import xml.etree.ElementTree as ET

tree = ET.parse('file.xml')

root = tree.getroot()

 

for neighbor in root.iter('B2B_METADATA'):

    for element in neighbor:

        if element.tag == 'EventId':

            print(element.text)

requested O/P: EventId 122157660

Answer 1

Actually for very simple queries, built-in etree supports limited XPath :

print(root.findall('.//B2B_METADATA/EventId')[0].text)

Similar to lxml's xpath :

print(root.xpath('//B2B_METADATA/EventId')[0].text)

Or by parsed objects:

print(root.find('B2B_METADATA').find('EventId').text)

Answer 2

To move your iterators down into XPath, you could use something like this:

result = tree.xpath('/B2B_DATA/B2B_METADATA/EventId/text()')

That would return a string representation of the text node contained in the EventId element (nested in a B2B_METADATA element, nested in a B2B_DATA element) in your XML, ie 122157660 . If there were multiple such text nodes in the XML then the xpath method will return them all as a list of strings.

If you knew that EventId only ever appears inside /B2B_DATA/B2B_METADATA then you could shorten your XPath to //EventId/text() . It would be computationally less efficient, because the // would search the entire document for EventId elements, but you may value conciseness over efficiency, especially if the XML document is really small (like your sample)