How to select parent based on the child in lxml?

问题

I have this code:

<table cellspacing="1" cellpadding="1" border="0">
  <tbody>
   <tr>
    <td>Something else</td>
   </tr>
   <tr>
    <td valign="top">
      <a href="http://exact url">Something</a>
    </td>
    <td valign="top">Something else</td>
   </tr>
  </tbody>
</table>

I want to find the Table but is very hard to target it (the very same code is used like 10 times). But I know what is in the URL. How can I get then the parent table?

回答1:

If t is the etree for this snippet of XML, then the link you're looking for is

t.xpath('//a[@href = "http://exact url"]')[0]

From there, you can get to the table using the ancestor axis:

t.xpath('//a[@href = "http://exact url"]/ancestor::table')[-1]

回答2:

Filter tables by using []. Note that the attribute is a grandchild //table[.//@href="blah"]

Or //a[@href="blah"]//ancestor::table

回答3:

A pure XPath solution.

Use:

(//a[@href = "http://exact url"])[1]/ancestor::table[1]

This selects the first ancestor table of the first a element in the XML document, the string value of whose href attribute is the string "http://exact url".

This provides the correct table element even in case when there are nested tables each of which has the wanted a element as descendant. In this case the above XPath expression selects the innermost such table -- in contrast with the currently accepted answer, that obtains the outermost table ancestor.