CODEWITHSUNDEEP
xmlperlsedawktext-processing
Ulrich Flamm
Ulrich Flamm

Reputation: 13

Extract data from xml file

I have a xml file containing thousands of entries, like:

<gml:featureMember>
<Feature>
<featureType>JCSOutput</featureType>
<property name="gml2_coordsys"></property>
<gml:PointProperty>
                <gml:Point>
                  <gml:coordinates>4048313.294966287,5374397.792158723 </gml:coordinates>
                </gml:Point>

</gml:PointProperty>
<property name="BEZEICHNUN">Anton-Bosch-Gasse</property>
<property name="WL_NUMMER">68</property>
</Feature>
</gml:featureMember>


<gml:featureMember>
<Feature>
<featureType>JCSOutput</featureType>
<property name="gml2_coordsys"></property>
<gml:PointProperty>
                <gml:Point>
                  <gml:coordinates>4044355.0231338665,5365146.95116724 </gml:coordinates>
                </gml:Point>

</gml:PointProperty>
<property name="BEZEICHNUN">Anschützgasse</property>
<property name="WL_NUMMER">67</property>
</Feature>
</gml:featureMember>

The script should search for a name given in a list (for example Anton-Bosch-Gasse) and copy the whole paragraph starting with <gml:featureMember> to a new file

What would you use for this purpose - awk, sed, perl?

Upvotes: 1

Views: 4295

Answers (3)

Joel Berger
Joel Berger

Reputation: 20280

Here is a solution like choroba's but using the Mojolicious suite. Its module Mojo::DOM traverses the XML using css3 selectors rather than xpath.

Here I find first all of the gml:featureMember elements, then extracts the first one which has a descendant that matches.

#!/usr/bin/env perl

use strict;
use warnings;

use Mojo::DOM;
use Mojo::Util qw/slurp spurt/;

my $dom = Mojo::DOM->new->xml(1);

# read in from file
# $dom->parse( slurp 'myfile.xml' );
# but for the demo ...
$dom->parse(do{ local $/; <DATA> });

my $found = 
  $dom->find('gml\:featureMember')
      ->first(sub{ 
        $_->find('property[name="BEZEICHNUN"]')
          ->first( qr/\QAnton-Bosch-Gasse/ )
      });

spurt "$found", 'output.xml';


__DATA__
<gml:featureMember>
<Feature>
<featureType>JCSOutput</featureType>
<property name="gml2_coordsys"></property>
<gml:PointProperty>
                <gml:Point>
                  <gml:coordinates>4048313.294966287,5374397.792158723 </gml:coordinates>
                </gml:Point>

</gml:PointProperty>
<property name="BEZEICHNUN">Anton-Bosch-Gasse</property>
<property name="WL_NUMMER">68</property>
</Feature>
</gml:featureMember>


<gml:featureMember>
<Feature>
<featureType>JCSOutput</featureType>
<property name="gml2_coordsys"></property>
<gml:PointProperty>
                <gml:Point>
                  <gml:coordinates>4044355.0231338665,5365146.95116724 </gml:coordinates>
                </gml:Point>

</gml:PointProperty>
<property name="BEZEICHNUN">Anschützgasse</property>
<property name="WL_NUMMER">67</property>
</Feature>
</gml:featureMember>

For this example I grab the XML from the DATA section. You might use the commented code to parse from a file.

You can also be a little more efficient if you are sure that the property is two deep in the structure consistently.

my $found = 
  $dom->find('gml\:featureMember property[name="BEZEICHNUN"]')
      ->first( qr/\QAnton-Bosch-Gasse/ )
      ->parent
      ->parent;

Upvotes: 1

mirod
mirod

Reputation: 16161

Using xml_grep, which comes with XML::Twig, you can write

$ xml_grep --root 'gml:featureMember' \
--cond 'property[string()="Anton-Bosch-Gasse"]' \
to_grep.xml > extract.xml

Upvotes: 3

choroba
choroba

Reputation: 241928

Sed and awk are not the right tools to parse XML. Reach for Perl:

#!/usr/bin/perl
use warnings;
use strict;

use XML::LibXML;

my $search = 'Anton-Bosch-Gasse';

# Put your real values here!
my $file = '1.xml';
my $uri  = 'http://1.2.3';

my $xpc = XML::LibXML::XPathContext->new;
$xpc->registerNs('gml', $uri);

my $xml = XML::LibXML->load_xml(location => $file);
my $r = $xml->find("//property[.='$search']/ancestor::gml:featureMember");
print $_->serialize for @$r;

Or, if you find the above example too verbose, you can use xsh:

my $search = 'Anton-Bosch-Gasse' ;
register-namespace gml http://1.2.3 ; # Insert the real URI.
open 1.xml ;                          # Insert the real path.
ls //property[.=$search]/ancestor::gml:featureMember ;

Upvotes: 5

Related Questions