From 0e4f36b653af8425650dc76130cfa25af50429ae Mon Sep 17 00:00:00 2001
From: Ere Maijala <ere.maijala@helsinki.fi>
Date: Thu, 8 Nov 2018 15:30:12 +0200
Subject: [PATCH] Bug 11529: Simplify and optimize batchRebuildBiblioTables.pl

Signed-off-by: Michal Denar <black23@gmail.com>
Signed-off-by: Michal Denar <black23@gmail.com>
Signed-off-by: Josef Moravec <josef.moravec@gmail.com>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Martin Renvoize <martin.renvoize@ptfs-europe.com>
---
 misc/batchRebuildBiblioTables.pl | 94 +++++++++++++-------------------
 1 file changed, 37 insertions(+), 57 deletions(-)

diff --git a/misc/batchRebuildBiblioTables.pl b/misc/batchRebuildBiblioTables.pl
index d66c4ce6db..15f650377a 100755
--- a/misc/batchRebuildBiblioTables.pl
+++ b/misc/batchRebuildBiblioTables.pl
@@ -15,90 +15,70 @@ BEGIN {
 # Koha modules used
 use Koha::Script;
 use MARC::Record;
+use C4::Charset;
 use C4::Context;
 use C4::Biblio;
 use Time::HiRes qw(gettimeofday);
 
 use Getopt::Long;
-my ( $input_marc_file, $number) = ('', 0);
-my ($version, $confirm, $test_parameter);
+
+my ($version, $confirm);
 GetOptions(
     'c' => \$confirm,
-    'h' => \$version,
-    't' => \$test_parameter,
+    'h' => \$version
 );
 
 if ($version || (!$confirm)) {
     print <<EOF
-This script rebuilds the non-MARC DB from the MARC values.
+This script rebuilds the non-MARC fields from the MARC values.
 You can/must use it when you change your mapping.
 
 Example: you decide to map biblio.title to 200\$a (it was previously mapped to 610\$a).
-Run this script or you will have strange results in OPAC !
+Run this script or you will have strange results in the UI!
 
 Syntax:
-\t./batchRebuildBiblioTables.pl -h (or without arguments => shows this screen)
-\t./batchRebuildBiblioTables.pl -c (c like confirm => rebuild non marc DB (may be long)
-\t-t => test only, change nothing in DB
+\t./batchRebuildBiblioTables.pl -h (or without arguments => show this screen)
+\t./batchRebuildBiblioTables.pl -c (c like confirm => rebuild non-MARC fields (may take long)
 EOF
 ;
     exit;
 }
 
+$|=1; # non-buffered output
+
 my $dbh = C4::Context->dbh;
 my $i=0;
-my $starttime = time();
-
-$|=1; # flushes output
-$starttime = gettimeofday;
+my $starttime = gettimeofday;
+my $marcflavour = C4::Context->preference('marcflavour');
+my $sth = $dbh->prepare('SELECT biblionumber, frameworkcode FROM biblio');
+$sth->execute();
 
-#1st of all, find item MARC tag.
-my ($tagfield,$tagsubfield) = &GetMarcFromKohaField( "items.itemnumber" );
-# $dbh->do("lock tables biblio write, biblioitems write, items write, marc_biblio write, marc_subfield_table write, marc_blob_subfield write, marc_word write, marc_subfield_structure write");
-my $sth = $dbh->prepare("SELECT biblionumber FROM biblio");
-$sth->execute;
-# my ($biblionumbermax) =  $sth->fetchrow;
-# warn "$biblionumbermax <<==";
 my @errors;
-while (my ($biblionumber)= $sth->fetchrow) {
-    #now, parse the record, extract the item fields, and store them in somewhere else.
-    my $record = GetMarcBiblio({ biblionumber => $biblionumber });
-    if (not defined $record) {
-	push @errors, $biblionumber;
-	next;
+while (my ($biblionumber, $frameworkcode) = $sth->fetchrow) {
+    my $marcxml = GetXmlBiblio($biblionumber);
+    if (not defined $marcxml) {
+        push @errors, $biblionumber;
+        next;
     }
-    my @fields = $record->field($tagfield);
-    my @items;
-    my $nbitems=0;
-    print ".";
-    my $timeneeded = gettimeofday - $starttime;
-    print "$i in $timeneeded s\n" unless ($i % 50);
-    $i++;
-    foreach my $field (@fields) {
-        my $item = MARC::Record->new();
-        $item->append_fields($field);
-        push @items,$item;
-        $record->delete_field($field);
-        $nbitems++;
+
+    $marcxml = C4::Charset::StripNonXmlChars( $marcxml );
+    my $record = eval {
+        MARC::Record::new_from_xml($marcxml, 'utf8', $marcflavour);
+    };
+    if ($@) {
+        push @errors, $biblionumber;
+        next;
     }
-#     print "$biblionumber\n";
-    my $frameworkcode = GetFrameworkCode($biblionumber);
-    localNEWmodbiblio($dbh,$record,$biblionumber,$frameworkcode) unless $test_parameter;
-}
-# $dbh->do("unlock tables");
-my $timeneeded = time() - $starttime;
-print "$i MARC record done in $timeneeded seconds\n";
-if (scalar(@errors) > 0) {
-    print "Some biblionumber could not be processed though: ", join(" ", @errors);
+
+    my $biblio = TransformMarcToKoha($record);
+    C4::Biblio::_koha_modify_biblio($dbh, $biblio, $frameworkcode);
+    C4::Biblio::_koha_modify_biblioitem_nonmarc($dbh, $biblio);
+
+    $i++;
+    printf("%lu records processed in %.2f seconds\n", $i, gettimeofday() - $starttime) unless ($i % 100);
 }
 
-# modified NEWmodbiblio to jump the MARC part of the biblio modif
-# highly faster
-sub localNEWmodbiblio {
-    my ($dbh,$record,$biblionumber,$frameworkcode) =@_;
-    $frameworkcode="" unless $frameworkcode;
-    my $oldbiblio = TransformMarcToKoha($record,$frameworkcode);
-    C4::Biblio::_koha_modify_biblio( $dbh, $oldbiblio, $frameworkcode );
-    C4::Biblio::_koha_modify_biblioitem_nonmarc( $dbh, $oldbiblio );
-    return 1;
+printf("\n%lu records processed in %.2f seconds\n", $i, gettimeofday() - $starttime);
+if (scalar(@errors) > 0) {
+    print "Some records could not be processed though: ", join(' ', @errors);
 }
-- 
2.39.5