Extracting a list of list with xpath

Question

I would like to use an xpath to get a list of list (or sequence of sequence) that groups extracted xml tags by parent element in order.

Here are my attempts so far using a minimal example..

import elementpath, lxml.etree
xml = '''<a>
<b c="1">
  <d e="3"/>
  <d e="4"/>
</b>
<b c="2">
  <d e="5"/>
  <d e="6"/>
</b>
</a>'''
tree = lxml.etree.fromstring(str.encode(xml))
xpath1 = '/a/b/d/@e'
xpath2 = 'for $b in (/a/b) return concat("[", $b/string-join(d/@e, ", "), "]")'
print('1:', elementpath.select(tree, xpath1))
print('2:', elementpath.select(tree, xpath2))
print('3:', [['3', '4'], ['5', '6']])

Which outputs..

1: ['3', '4', '5', '6']
2: ['[3, 4]', '[5, 6]']
3: [['3', '4'], ['5', '6']]

xpath1 returns a flattened list/sequence, with no grouping by parent element.

xpath2 is the closest I have come so far, but gives sub-arrays as string rather than array.

option 3 is what I am after

Anyone able to advise on a better way of doing this with just an xpath?

Thanks, Mark

Is it a requirement that the result comes directly from the xpath expression? Because [x.xpath('d/@e') for x in tree.xpath('/a/b')] gives you what you want. — larsks
– larsks, Commented Jul 19, 2023 at 2:31
thanks for the suggestion.. i am using xpaths to configure an xml parser, and would like to avoid python code in the parser configuration data — Mark
– Mark, Commented Jul 19, 2023 at 4:27

Martin Honnen · Accepted Answer · 2023-07-19 13:30:17Z

ElementPath supports XPath 3.1 with XPath/XDM arrays so I think you want, in terms of XPath

/a!array { b ! array { d/@e/string() } }

which should give [["3","4"],["5","6"]].

That is the output with SaxonC HE (12.3) of

from saxonche import PySaxonProcessor

xml = '''<a>
<b c="1">
  <d e="3"/>
  <d e="4"/>
</b>
<b c="2">
  <d e="5"/>
  <d e="6"/>
</b>
</a>'''

with PySaxonProcessor(license=False) as saxon:
    xdm_doc = saxon.parse_xml(xml_text=xml)
    xpath_processor = saxon.new_xpath_processor()
    xpath_processor.set_context(xdm_item=xdm_doc)
    xdm_value = xpath_processor.evaluate_single('/a!array { b ! array { d/@e/string() } }')
    print(xdm_value)

At that stage you don't have a Python list of lists, however, rather a PyXdmItem which is an XDM array of arrays, to get a nested Python list I think you can do

    list_of_lists = [inner_array.head.as_list() for inner_array in xdm_array.as_list()]
    print(list_of_lists)

I will need to check whether ElementPath allows that too and perhaps a bit more elegantly; the simplest I have found is

import elementpath, lxml.etree
from elementpath.xpath3 import XPath3Parser

xml = '''<a>
<b c="1">
  <d e="3"/>
  <d e="4"/>
</b>
<b c="2">
  <d e="5"/>
  <d e="6"/>
</b>
</a>'''

tree = lxml.etree.fromstring(str.encode(xml))

array_of_arrays = elementpath.select(tree, '/a!array { b ! array { d/@e/string() } }', parser=XPath3Parser)

print(array_of_arrays)

list_of_lists = [array.items() for array in array_of_arrays[0].items()]

print(list_of_lists)

giving [['3', '4'], ['5', '6']] for the final pint(list_of_lists).

Or using a sequence of arrays in XPath gives you a list of arrays in Python which you can more easily convert into a list of lists in Python:

sequence_of_arrays = elementpath.select(tree, '/a/b ! array { d/@e/string() }', parser=XPath3Parser)

print(sequence_of_arrays)

list_of_lists = [array.items() for array in sequence_of_arrays]

print(list_of_lists)

Collectives™ on Stack Overflow

Extracting a list of list with xpath

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related