From 1c92a0dabca329d4a876e90819d8e7be8494140e Mon Sep 17 00:00:00 2001
From: Jason Etheridge <jason@esilibrary.com>
Date: Thu, 7 Aug 2008 18:27:14 +0000
Subject: [PATCH] select marc and create one text file per selection

---
 select_marc_as_text.pl |   55 ++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 55 insertions(+), 0 deletions(-)
 create mode 100755 select_marc_as_text.pl

diff --git a/select_marc_as_text.pl b/select_marc_as_text.pl
new file mode 100755
index 0000000..955edd9
--- /dev/null
+++ b/select_marc_as_text.pl
@@ -0,0 +1,55 @@
+#!/usr/bin/perl
+use open ':utf8';
+use MARC::Batch;
+use MARC::Record;
+use MARC::File::XML ( BinaryEncoding => 'utf-8' );
+use MARC::Field;
+
+my $record_id_file = $ARGV[0];
+my %record_ids;
+
+open FILE, $record_id_file;
+while (my $record_id = <FILE>) {
+    chomp($record_id); $record_ids{ $record_id } = 1;
+}
+close FILE;
+
+my $id_tag = $ARGV[1]; my $id_subfield = $ARGV[2];
+
+binmode(STDOUT, ':utf8');
+binmode(STDIN, ':utf8');
+
+my $M;
+
+foreach $argnum ( 3 .. $#ARGV ) {
+
+	print STDERR "Processing " . $ARGV[$argnum] . "\n";
+
+    open $M, '<:utf8', $ARGV[$argnum];
+
+	my $batch = MARC::Batch->new('XML',$M);
+	$batch->strict_off();
+	$batch->warnings_off();
+
+    my $count = 0;
+
+	while ( my $record = $batch->next() ) {
+
+        $count++;
+
+		my $id = $record->field($id_tag);
+		if (!$id) {
+			print STDERR "ERROR: This record is missing a $id_tag field.\n" . $record->as_formatted() . "\n=====\n";
+			next;
+		}
+		$id = $id->as_string($id_subfield);
+
+        if (defined $record_ids{ $id }) {
+            open FILE, ">$id.txt";
+            binmode(FILE, ':utf8');
+            print FILE $record->as_formatted();
+            close FILE;
+        }
+	}
+    print STDERR "Processed $count records.\n";
+}
-- 
1.7.2.5