#!/usr/bin/perl

# This example reads the output.xml file generated by pacific.pl. You
# certainly want to run that one first. With slight modifications, it
# can read other files as well - but the multi-character entity decoding
# won't be obvious (and might had adverse side-effects). Many of the
# same comments apply to this example. It reads incrementally to allow
# handling large records.

# First, a few administrative details. If you run this from someplace
# other than the "eg" directory of the distribution, you will probably
# need to fix the paths.

    use lib '../blib/lib';
    use MARC::XML 0.3;

    unlink "output2.txt";
    my $m = MARC::XML->new();

# We'll create a couple of "identifiable" multi-character-output entities.
# We start with the "base" set - although a completely independent set is
# also possible. You could even do UCS-2 or UTF-8 this way. But unless you
# already have the rest of the code in those formats, this translation
# will not be a complete solution. And for UTF-8, a more efficient option
# is possible by using a different DTD file to redefine entities in the
# "internal" character map used by Expat.

    my $charhash = $m->register_default;	# MARC 8-bit character set
    ${$charhash}{softsign} = '[I am the VERY model]';
    ${$charhash}{caron} = '[of a VERY modern entity]';

# Setup an XML::Parser and define options.

    my $count = $m->openxml({file=>"output.xml",charset=>$charhash});

    while ($m->nextxml(1)) {

# Loop printing titles.

        my ($title) = $m->getvalue({record=>'1',field=>'245',subfield=>'a'});
	print "$title\n";

# Use a different name for the text output than used by pacific.pl.
# Hence, you can compare them.

	$m->output({file=>">>output2.txt",format=>"ascii"});
        $m->deletemarc(); #empty the object for reading in another
	$count++;
    }
    $m->closexml || die "Can't close";

    print "\nprocessed $count records\n";