bulkmarcimport.pl - Import bibliographic/authority records into Koha
$ export KOHA_CONF=/etc/koha.conf
$ perl misc/migration_tools/bulkmarcimport.pl -d --commit 1000 \\
--file /home/jmf/koha.mrc -n 3000
Don't use this script before you've entered and checked your MARC parameters tables twice (or more!). Otherwise, the import won't work correctly and you will get invalid data.
This version/help screen
Type of import: bibliographic records
Type of import: authority records
The FILE to import
Verbose mode. 1 means "some infos", 2 means "MARC dumping"
Turn off foreign key checks during import.
The NUMBER of records to import. If missing, all the file is imported
File offset before importing, ie NUMBER of records to skip.
The NUMBER of records to wait before performing a 'commit' operation
File logs actions done for each record and their status into file
If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
Test mode: parses the file, saying what it would do, but doing nothing.
Skip automatic conversion of MARC-8 to UTF-8. This option is provided for debugging.
The CHARACTERISTIC MARC flavour. At the moment, only MARC21 and UNIMARC are supported. MARC21 by default.
Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio, biblioitems, items
Input file FORMAT: MARCXML or ISO2709 (defaults to ISO2709)
file yamlfile with authoritiesTypes and distinguishable record field in order to store the correct authtype
yaml file format a yaml file with ids
list of fields that will not be imported. Can be any from 000 to 999 or field, subfield and subfield's matching value such as 200avalue
if set, only insert when possible
if set, only updates (any biblio should have a matching record)
if set, do whatever is required
Field store ids in FIELD (useful for authorities, where 001 contains the authid for Koha, that can contain a very valuable info for authorities coming from LOC or BNF. useless for biblios probably)
FIELD matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch can be either 001 to 999 or field and list of subfields as such 100abcde
If set, a search will be done on isbn, and, if the same isbn is found, the biblio is not added. It's another method to deduplicate. -match & -isbn can be both set.
Clean ISBN fields from entering biblio records, ie removes hyphens. By default, ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
Source bib TAG for reporting the source bib number
Source SUBFIELD for reporting the source bib number
FILE for the koha bib and source id
Store ids in 009 (useful for authorities, where 001 contains the authid for Koha, that can contain a very valuable info for authorities coming from LOC or BNF. useless for biblios probably)
If set, whenever a duplicate barcode is detected, it is removed and the attempt to add the record is retried, thereby giving the record a blank barcode. This is useful when something has set barcodes to be a biblio ID, or similar (usually other software.)
This is the code for the framework that the requested records will have attached to them when they are created. If not specified, then the default framework will be used.
This parameter allows you to use a local module with a customize subroutine that is called for each MARC record. If no filename is passed, LocalChanges.pm is assumed to be in the migration_tools subdirectory. You may pass an absolute file name or a file name from the migration_tools directory.
This parameter allows you to specify the name of an existing MARC modification template to apply as the MARC records are imported (these templates are created in the "MARC modification templates" tool in Koha). If not specified, no MARC modification templates are used (default).
If set, do not index the imported records with Zebra or Elasticsearch. Use this when you plan to do a complete reindex of your data after running bulkmarciport. This can increase performance and avoid unnecessary load.
If set, check the validity of records before adding. If they are invalid we will print the output of MARC::Lint->check_record and skip them during the import. Without this option bad records may kill the job.