++ed by:

23 PAUSE users
19 non-PAUSE users.

Michel Rodriguez


xml_grep - grep XML files looking for specific elements


  xml_grep [options] <file list> 


  xml_grep <xpath expression> <file list>

By default you can just give xml_grep an XPath expression and a list of files, and get an XML file with the result.

This is equivalent to writing

  xml_grep --group_by_file file --pretty_print indented --cond <file list>



brief help message


full documentation


display the tool version

--root <cond>

look for and return xml chunks matching <cond>

if neither --root nor --file are used then the element(s) that trigger the --cond option is (are) used. If --cond is not used then all elements matching the <cond> are returned

several --root can be provided

--cond <cond>

return the chunks (or file names) only if they contain elements matching <cond>

several --cond can be provided (in which case they are OR'ed)


return only file names (do not generate an XML output)

usage of this option precludes using any of the options that define the XML output: --roots, --encoding, --wrap, --group_by_file or --pretty_print


return only the number of matches in each file

usage of this option precludes using any of the options that define the XML output: --roots, --encoding, --wrap, --group_by_file or --pretty_print


without this option parsing errors are reported to STDOUT and the file skipped


when on (by default) the wrapping element get a date attribute that gives the date the tool was run.

with --nodate this attribute is not added, which can be useful if you need to compare 2 runs.

--encoding <enc>

encoding of the xml output (utf-8 by default)

--nb_results <nb>

output only <nb> results


output only <nb> results by file

--wrap <tag>

wrap the xml result in the provided tag (defaults to 'xml_grep')

If wrap is set to an empty string (--wrap '') then the xml result is not wrapped at all.


same as using --wrap '': the xml result is not wrapped.

--descr <string>

attributes of the wrap tag (defaults to version="<VERSION>" date="<date>")

--group_by_file <optional_tag>

wrap results for each files into a separate element. By default that element is named file. It has an attribute named filename that gives the name of the file.

the short version of this option is -g

--exclude <condition>

same as using -v in grep: the elements that match the condition are excluded from the result, the input file(s) is (are) otherwise unchanged

the short form of this option is -v

--pretty_print <optional_style>

pretty print the output using XML::Twig styles ('indented', 'record' or 'record_c' are probably what you are looking for)

if the option is used but no style is given then 'indented' is used

short form for this argument is -s


Displays the text of the results, one by line.


Allow HTML input, files are converted using HTML::TreeBuilder


Allow HTML input, files are converted using HTML::Tidy

Condition Syntax

<cond> is an XPath-like expression as allowed by XML::Twig to trigger handlers.

exemples: 'para' 'para[@compact="compact"]' '*[@urgent]' '*[@urgent="1"]' 'para[string()="WARNING"]'

see XML::Twig for a more complete description of the <cond> syntax

options are processedby Getopt::Long so they can start with '-' or '--' and can be abbreviated (-r instead of --root for example)


xml_grep does a grep on XML files. Instead of using regular expressions it uses XPath expressions (in fact the subset of XPath supported by XML::Twig)

the results can be the names of the files or XML elements containing matching elements.


XML::Twig Getopt::Long


This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.


Michel Rodriguez <mirod@xmltwig.com>