2 # small script that import an iso2709 file into koha 2.0
7 use MARC::File::USMARC;
12 use Time::HiRes qw(gettimeofday);
15 my ( $input_marc_file, $number) = ('',0);
16 my ($version, $delete, $test_parameter,$char_encoding, $verbose);
18 'file:s' => \$input_marc_file,
22 't' => \$test_parameter,
23 'c:s' => \$char_encoding,
27 if ($version || ($input_marc_file eq '')) {
29 small script to import an iso2709 file into Koha.
31 \th : this version/help screen
32 \tfile /path/to/file/to/dump : the file to dump
33 \tv : verbose mode. 1 means "some infos", 2 means "MARC dumping"
34 \tn : the number of the record to import. If missing, all the file is imported
35 \tt : test mode : parses the file, saying what he would do, but doing nothing.
36 \tc : the char encoding. At the moment, only MARC21 and UNIMARC supported. MARC21 by default.
37 \d : delete EVERYTHING related to biblio in koha-DB before import :tables :
38 \t\tbiblio, \t\tbiblioitems, \t\tsubjects,\titems
39 \t\tadditionalauthors, \tbibliosubtitles, \tmarc_biblio,
40 \t\tmarc_subfield_table, \tmarc_word, \t\tmarc_blob_subfield
41 IMPORTANT : don't use this script before you've entered and checked twice (or more) your MARC parameters tables.
42 If you fail this, the import won't work correctly and you will get invalid datas.
44 SAMPLE : ./bulkmarcimport.pl -file /home/paul/koha.dev/local/npl -n 1
50 my $dbh = C4::Context->dbh;
53 print "deleting biblios\n";
54 $dbh->do("delete from biblio");
55 $dbh->do("delete from biblioitems");
56 $dbh->do("delete from items");
57 $dbh->do("delete from bibliosubject");
58 $dbh->do("delete from additionalauthors");
59 $dbh->do("delete from bibliosubtitle");
60 $dbh->do("delete from marc_biblio");
61 $dbh->do("delete from marc_subfield_table");
62 $dbh->do("delete from marc_word");
63 $dbh->do("delete from marc_blob_subfield");
65 if ($test_parameter) {
66 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
69 $char_encoding = 'MARC21' unless ($char_encoding);
70 print "CHAR : $char_encoding\n" if $verbose;
71 my $starttime = gettimeofday;
72 my $batch = MARC::Batch->new( 'USMARC', $input_marc_file );
73 $batch->warnings_off();
76 #1st of all, find item MARC tag.
77 my ($tagfield,$tagsubfield) = &MARCfind_marc_from_kohafield($dbh,"items.itemnumber");
78 # $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, stopwords write");
79 while ( my $record = $batch->next() ) {
81 #now, parse the record, extract the item fields, and store them in somewhere else.
83 ## create an empty record object to populate
84 my $newRecord = MARC::Record->new();
86 # go through each field in the existing record
87 foreach my $oldField ( $record->fields() ) {
89 # just reproduce tags < 010 in our new record
90 if ( $oldField->tag() < 10 ) {
91 $newRecord->append_fields( $oldField );
95 # store our new subfield data in this list
96 my @newSubfields = ();
98 # go through each subfield code/data pair
99 foreach my $pair ( $oldField->subfields() ) {
100 $pair->[1] =~ s/\<//g;
101 $pair->[1] =~ s/\>//g;
102 push( @newSubfields, $pair->[0], char_decode($pair->[1],$char_encoding) );
105 # add the new field to our new record
106 my $newField = MARC::Field->new(
108 $oldField->indicator(1),
109 $oldField->indicator(2),
113 $newRecord->append_fields( $newField );
118 warn "$i ==>".$newRecord->as_formatted() if $verbose eq 2;
119 my @fields = $newRecord->field($tagfield);
123 foreach my $field (@fields) {
124 my $item = MARC::Record->new();
125 $item->append_fields($field);
127 $record->delete_field($field);
130 print "$i : $nbitems items found\n" if $verbose;
131 # now, create biblio and items with NEWnewXX call.
132 unless ($test_parameter) {
133 my ($bibid,$oldbibnum,$oldbibitemnum) = NEWnewbiblio($dbh,$newRecord);
134 warn "ADDED biblio NB $bibid in DB\n" if $verbose;
135 for (my $i=0;$i<=$#items;$i++) {
136 NEWnewitem($dbh,$items[$i],$bibid);
140 # $dbh->do("unlock tables");
141 my $timeneeded = gettimeofday - $starttime;
142 print "$i MARC record done in $timeneeded seconds";