pymarc 5.4.0

Read, write and modify MARC bibliographic data

_|_|_|    _|    _|  _|_|_|  _|_|      _|_|_|  _|  _|_|    _|_|_|
_|    _|  _|    _|  _|    _|    _|  _|    _|  _|_|      _|
_|    _|  _|    _|  _|    _|    _|  _|    _|  _|        _|
_|_|_|      _|_|_|  _|    _|    _|    _|_|_|  _|          _|_|_|
_|              _|
_|          _|_|

pymarc is a python library for working with bibliographic data encoded in MARC21. It provides an API for reading, writing and modifying MARC records. It was mostly designed to be an emergency eject seat, for getting your data assets out of MARC and into some kind of saner representation. However over the years it has been used to create and modify MARC records, since despite repeated calls for it to die as a format, MARC seems to be living quite happily as a zombie.

Below are some common examples of how you might want to use pymarc. If you run across an example that you think should be here please send a pull request.

You can read pymarc documentation here.

Installation

You'll probably just want to use pip to install pymarc:

pip install pymarc

Reading

Most often you will have some MARC data and will want to extract data from it. Here's an example of reading a batch of records and printing out the title. If you are curious this example uses the batch file available here in pymarc repository:

from pymarc import MARCReader
with open('test/marc.dat', 'rb') as fh:
    reader = MARCReader(fh)
    for record in reader:
        print(record.title)

The pragmatic programmer : from journeyman to master /
Programming Python /
Learning Python /
Python cookbook /
Python programming for the absolute beginner /
Web programming : techniques for integrating Python, Linux, Apache, and MySQL /
Python programming on Win32 /
Python programming : an introduction to computer science /
Python Web programming /
Core python programming /
Python and Tkinter programming /
Game programming with Python, Lua, and Ruby /
Python programming patterns /
Python programming with the Java class libraries : a tutorial for building Web
and Enterprise applications /
Learn to program using Python : a tutorial for hobbyists, self-starters, and all
who want to learn the art of computer programming /
Programming with Python /
BSD Sockets programming from a multi-language perspective /
Design patterns : elements of reusable object-oriented software /
Introduction to algorithms /
ANSI Common Lisp /

A pymarc.Record object has a few handy properties like title for getting at bits of a bibliographic record, others include: author, isbn, subjects, location, notes, physicaldescription, publisher, pubyear, issn, issn_title. But really, to work with MARC data you need to understand the numeric field tags and subfield codes that are used to designate various bits of information. There is a lot more data hidden in a MARC record than these helper properties provide access to. For example the title property works by extracting the information from the 245 field, subfields a and b behind the scenes. You can access 245a like so:

print(record['245']['a'])

Some fields like subjects can repeat. In cases like that you will want to use get_fields to get all of them as pymarc.Field objects, which you can then interact with further:

for f in record.get_fields('650'):
    print(f)

If you are new to MARC fields Understanding MARC is a pretty good primer, and the MARC 21 Formats page at the Library of Congress is a good reference once you understand the basics.

Writing

Note: As of v5.0.0 Subfield is used to create subfields. Prior to v5, subfields were constructed and accessed as a list of strings, e.g., [code, value, code, value]. In v5.0.0 this has been changed to organize the subfields into a list of tuples, e.g., [(code, value), (code, value)]. The Subfield is implemented as a NamedTuple so that the tuples can be constructed as Subfield(code=code, value=value). The old style of creating subfields is no longer supported. Attempting to pass a list of strings to the subfields parameter for the Field constructor will raise a ValueError. For convenience the Field.convert_legacy_subfields class method can be used to convert a legacy list of strings into a list of Subfields.

Here's an example of creating a record and writing it out to a file.

from pymarc import Record, Field, Subfield, Indicators

record = Record()
record.add_field(
    Field(
        tag='245',
        indicators=Indicators('0', '1'),
        subfields=[
            Subfield(code='a', value='The pragmatic programmer : '),
            Subfield(code='b', value='from journeyman to master /'),
            Subfield(code='c', value='Andrew Hunt, David Thomas.')
        ]))
with open('file.dat', 'wb') as out:
    out.write(record.as_marc())

To convert from the old string list to a list of Subfields, the .convert_legacy_subfields class method is provided on the Field class.

from pymarc import Field, Subfield

legacy_fields: list[str] = ['a', 'The pragmatic programmer : ',
                            'b', 'from journeyman to master /',
                            'c', 'Andrew Hunt, David Thomas']

coded_fields: list[Subfield] = Field.convert_legacy_subfields(legacy_fields)

Updating

Updating works the same way, you read it in, modify it, and then write it out again:

from pymarc import MARCReader
with open('test/marc.dat', 'rb') as fh:
    reader = MARCReader(fh)
    record = next(reader)
    record['245']['a'] = 'The Zombie Programmer : '
with open('file.dat', 'wb') as out:
    out.write(record.as_marc())

JSON and XML

If you find yourself using MARC data a fair bit, and distributing it, you may make other developers a bit happier by using the JSON or XML serializations. The main benefit to using XML or JSON is that the UTF8 character encoding is used, rather than the frustratingly archaic MARC8 encoding. Also they will be able to use standard JSON and XML reading/writing tools to get at the data they want instead of some crazy MARC processing library like, ahem, pymarc.

XML

To parse a file of MARCXML records you can:

from pymarc import parse_xml_to_array

records = parse_xml_to_array('test/batch.xml')

If you have a large XML file and would rather not read them all into memory you can:

from pymarc import map_xml

def print_title(r):
    print(r.title)

map_xml(print_title, 'test/batch.xml')

Also, if you prefer you can pass in a file like object in addition to the path to both map_xml and parse_xml_to_array:

from pymarc import parse_xml_to_array

records = parse_xml_to_array(open('test/batch.xml'))

JSON

JSON support is fairly minimal in that you can call a pymarc.Record's as_json() method to return JSON for a given MARC Record:

from pymarc import MARCReader

with open('test/one.dat','rb') as fh:
    reader = MARCReader(fh)
    for record in reader:
        print(record.as_json(indent=2))

{
  "leader": "01060cam  22002894a 4500",
  "fields": [
    {
      "001": "11778504"
    },
    {
      "010": {
        "ind1": " ",
        "subfields": [
          {
            "a": "   99043581 "
          }
        ],
        "ind2": " "
      }
    },
    {
      "100": {
        "ind1": "1",
        "subfields": [
          {
            "a": "Hunt, Andrew,"
          },
          {
            "d": "1964-"
          }
        ],
        "ind2": " "
      }
    },
    {
      "245": {
        "ind1": "1",
        "subfields": [
          {
            "a": "The pragmatic programmer :"
          },
          {
            "b": "from journeyman to master /"
          },
          {
            "c": "Andrew Hunt, David Thomas."
          }
        ],
        "ind2": "4"
      }
    },
    {
      "260": {
        "ind1": " ",
        "subfields": [
          {
            "a": "Reading, Mass :"
          },
          {
            "b": "Addison-Wesley,"
          },
          {
            "c": "2000."
          }
        ],
        "ind2": " "
      }
    },
    {
      "300": {
        "ind1": " ",
        "subfields": [
          {
            "a": "xxiv, 321 p. ;"
          },
          {
            "c": "24 cm."
          }
        ],
        "ind2": " "
      }
    },
    {
      "504": {
        "ind1": " ",
        "subfields": [
          {
            "a": "Includes bibliographical references."
          }
        ],
        "ind2": " "
      }
    },
    {
      "650": {
        "ind1": " ",
        "subfields": [
          {
            "a": "Computer programming."
          }
        ],
        "ind2": "0"
      }
    },
    {
      "700": {
        "ind1": "1",
        "subfields": [
          {
            "a": "Thomas, David,"
          },
          {
            "d": "1956-"
          }
        ],
        "ind2": " "
      }
    }
  ]
}

If you want to parse a file of MARCJSON records you can:

from pymarc import parse_json_to_array

records = parse_json_to_array(open('test/batch.json'))

print(records[0])

=LDR  00925njm  22002777a 4500
=001  5637241
=003  DLC
=005  19920826084036.0
=007  sdubumennmplu
=008  910926s1957\\\\nyuuun\\\\\\\\\\\\\\eng\\
=010  \\$a   91758335
=028  00$a1259$bAtlantic
=040  \\$aDLC$cDLC
=050  00$aAtlantic 1259
=245  04$aThe Great Ray Charles$h[sound recording].
=260  \\$aNew York, N.Y. :$bAtlantic,$c[1957?]
=300  \\$a1 sound disc :$banalog, 33 1/3 rpm ;$c12 in.
=511  0\$aRay Charles, piano & celeste.
=505  0\$aThe Ray -- My melancholy baby -- Black coffee -- There's no you -- Doodlin' -- Sweet sixteen bars -- I surrender dear -- Undecided.
=500  \\$aBrief record.
=650  \0$aJazz$y1951-1960.
=650  \0$aPiano with jazz ensemble.
=700  1\$aCharles, Ray,$d1930-$4prf

Developing

If you'd like to further develop pymarc you'll want to get the latest code:

git clone git://gitlab.com/pymarc/pymarc.git
cd pymarc

Install uv, and run the tests:

uv run pytest

Run the lint tests:

uv run ruff check .

Reformat any code that needs it:

uv run ruff format .

And run type checking:

uv run pyright .

If you want to update the dependencies:

uv lock --upgrade

If you want to build and publish a new version of pymarc on PyPI you can:

uv build
uv publish

Support

The pymarc developers encourage you to join the pymarc Google Group if you need help. Also, please feel free to use issue tracking on GitLab to submit feature requests or bug reports. If you've got an itch to scratch, please scratch it, and send merge requests on GitLab.

If you start working with MARC you may feel like you need moral support in addition to technical support. The #python channel on [code4lib Slack]https://code4lib.org/slack/) is a good place to start.