Bug #2373

RCP: 0.7.9, XTZ + CSV import module: error in page indexing if source file name contains the "_" character

Added by Alexey Lavrentev about 5 years ago. Updated almost 2 years ago.

Status:New Start date:04/30/2018
Priority:Normal Due date:
Assignee:- % Done:


Category:Import Spent time: -
Target version:TXM 0.8.2


If the corpus source directory contains files differentiated by underscore-separated suffixes, the index of pages in the import contains duplicates.


Sources files :
  • mytext.xml
  • mytext_a.xml

import.xml in the binary corpus:

            <text name="mytext">
               <source file="/home/user/TXM/corpora/MYCORPUS/txm/MYCORPUS/mytext.xml" 
                  <edition index="/home/user/TXM/corpora/MYCORPUS/HTML/MYCORPUS/default" 
                     <page id="1" wordid="w_0"/>
                     <page id="a_1" wordid="w_0"/>
                     <page id="a_2" wordid="mytext_a_1"/>
                     <page id="2" wordid="mytext_1"/>
                     <page id="3" wordid="mytext_137"/>
                     <page id="a_3" wordid="mytext_a_18"/>


  • Correct the regexp pattern when searching for pages to index
  • Use a more solid mechanism for page indexing

Temporary workaround

  • Document the restriction on file names
  • Patch binary corpus with an XSLT


#1 Updated by Alexey Lavrentev about 5 years ago

  • Target version changed from Portal 0.8 to TXM 0.8.0a (split/restructuration)

#2 Updated by Sebastien Jacquot almost 5 years ago

  • Target version changed from TXM 0.8.0a (split/restructuration) to TXM 0.8.0

#3 Updated by Matthieu Decorde about 4 years ago

  • Target version changed from TXM 0.8.0 to TXM 0.8.2

#4 Updated by Matthieu Decorde almost 3 years ago

  • Category set to Import

#5 Updated by Matthieu Decorde almost 2 years ago

  • % Done changed from 0 to 80

fixed with the new TXMResult objects (CorpusBuild, Text, Edition)

Also available in: Atom PDF