I have an xml-file with a format similar to docx, i.e.:
<w:r> <w:rPr> <w:sz w:val="36"/> <w:szCs w:val="36"/> </w:rPr> <w:t>BIG_TEXT</w:t> </w:r>
I need to get an index of BIG_TEXT
in source xml, like:
from lxml import etree text = open('/devel/tmp/doc2/word/document.xml', 'r').read() root = etree.XML(text) start = 0 for e in root.iter("*"): if e.text: offset = text.index(e.text, start) l = len(e.text) print 'Text "%s" at offset %s and len=%s' % (e.text, offset, l) start = offset + l
I can start a new search from position of current index
+ len(text)
, but is there another way? Element may have one character, w
for example. It will find index of w
, but not index of tag text w
.
Advertisement
Answer
I was looking for a similar solution (indexing nodes in a big xml file for fast lookup).
- AFAIK,
lxml
only offers sourceline, which is insufficient. Cf API :Original line number as found by the parser or None if unknown.
- But
expat
provides the exact offset in the file : CurrentByteIndex.- Fetched from
start_element
handler, it returns tag’s start (ie'<'
) offset. - Fetched from
char_data
handler, it returns data’s start (ie'B'
in your example) offset.
- Fetched from
Example :
import xml.parsers.expat # handler functions for parser events, and housekeeping. class handler : def __init__(self, current_parser) : #tag of interest self.TARGET_TAG = "w:t" #set up parser self.parser = current_parser self.parser.StartElementHandler = self.start_element self.parser.EndElementHandler = self.end_element self.parser.CharacterDataHandler = self.char_data self.target_tag_met = False self.index = None def start_element(self, name, attrs): self.target_tag_met = (name == self.TARGET_TAG) def end_element(self, name) : self.target_tag_met = False def char_data(self, data): if self.target_tag_met : self.index = self.parser.CurrentByteIndex #open file in binary mode for robuster byte offsets. xmlFile = open("so_test.xml", 'rb') p = xml.parsers.expat.ParserCreate() h = handler(p) p.ParseFile(xmlFile) print (h.index)