From: Don McMorris <dmcmorris@esilibrary.com>
Date: Wed, 30 Jul 2008 23:38:54 +0000 (+0000)
Subject: ci split_marc.pl
X-Git-Url: http://git.equinoxoli.org/?p=migration-tools.git;a=commitdiff_plain;h=92214de6bc40fc23a8a69d73d2e942c5ede13e9a

ci split_marc.pl
---

diff --git a/split_marc.pl b/split_marc.pl
new file mode 100644
index 0000000..f98eada
--- /dev/null
+++ b/split_marc.pl
@@ -0,0 +1,35 @@
+#!/usr/bin/perl
+use MARC::Batch;
+use MARC::Record;
+use MARC::File::XML ( BinaryEncoding => 'utf-8' );
+use MARC::Field;
+use POSIX;
+
+my $split_every = $ARGV[0];
+my $count = 0;
+
+binmode(STDOUT, ':utf8');
+binmode(STDIN, ':utf8');
+
+foreach $argnum ( 1 .. $#ARGV ) {
+
+        print STDERR "Processing " . $ARGV[$argnum] . "\n";
+
+        my $batch = MARC::Batch->new('XML',$ARGV[$argnum]);
+        $batch->strict_off();
+        $batch->warnings_off();
+
+        while ( my $record = $batch->next() ) {
+
+        $count++;
+
+                my $filename = $ARGV[$argnum] . ".split." .  floor( $count / $split_every ) . ".xml";
+
+                open FILE, ">>$filename";
+                binmode(FILE, ':utf8');
+                print FILE $record->as_xml();
+                close FILE;
+        }
+    print STDERR "Processed $count records.\n";
+}
+