Question

I wonder if it’s possible to make conditional statments connected with a tree.findall(“…”) statement in lxml library?

I have the following xml structure in a file

    <sss version="1.2">
    <date>2011-09-23</date>
    <time>12:32:29</time>
    <origin>OPST</origin>
    <user></user>
    <survey>
        <name>Test</name>
        <version>2011-09-02 15:50:10</version>
        <record ident="A">
            <variable ident="10" type="quantity">
                <name>no_v</name>
                <label>Another question</label>
                <position start="23" finish="24"/>
                <values>
                    <range from="0" to="32"/>
                </values>
            </variable>
            <variable ident="11" type="quantity">
                <name>v_683</name>
                <label>another totally another Question</label>
                <position start="25" finish="26"/>
                <values>
                    <range from="0" to="33"/>
                </values>
            </variable>
            <variable ident="12" type="quantity">
                <name>v_684</name>
                <label>And once more Question</label>
                <position start="27" finish="29"/>
                <values>
                    <range from="0" to="122"/>
                </values>
            </variable>
            <variable ident="20" type="single">
                <name>v_684</name>
                <label>Question with alternatives</label>
                <position start="73" finish="73"/>
                <values>
                    <range from="1" to="6"/>
                    <value code="1">Alternative 1</value>
                    <value code="2">Alternative 2</value>
                    <value code="3">Alternative 3</value>
                    <value code="6">Alternative 4</value>
                </values>
            </variable>
        </record>
    </survey>
</sss>

What I want to do now is to get only survey/record/variable/name .text and survey/record/variable/values/value .text if the name starts with “v_”

So far i have the first part

from lxml import etree as ET
tree = ET.parse('scheme.xml')
[elem.text for elem in tree.getiterator(tag='name') if elem.text.startswith('v_')]

But how can I get the survey/record/variable/values/value .text of the SAME element…and use survey/record/variable/name .text like a filter?
Thanks a lot!

Answer 1

[(elem.text,elem.getparent().xpath('values/value/text()')) 
 for elem in tree.getiterator(tag='name') if elem.text.startswith('v_')]

yields

[('v_683', []),
 ('v_684', []),
 ('v_684',
  ['Alternative 1', 'Alternative 2', 'Alternative 3', 'Alternative 4'])]

elem is a name element. So to get the the associated values, you can first find its parent (variable), then search for the values child, and then the value subchild elements.

---

An alternative that removes the getparent call, but uses a slightly more complicated XPath is:

[(elem.text,elem.xpath('following-sibling::values/value/text()')) for elem in tree.getiterator(tag='name')  if elem.text.startswith('v_')]

following-sibling:: tells xpath to generate all siblings of name.

following-sibling::values tells xpath to generate all siblings of name that are values elements.