Python script for converting MARC21 files to a saner format (CSV), originally designed for the freely available Harvard Library Bibliographic Dataset.
This script only conserves the following subset of MARC21 records (with additional simple post-processing on them):
| Field | Description |
|---|---|
| 020a | ISBN |
| 245a/245b | Title |
| 100/110/111 | Author |
| 260b | Publisher |
| 260a | Publisher Place |
| 300a | Extent (eg. pages) |
| 300c | Physical dimensions |
| 650c | Subject |
| All 5xx | Notes |
| 988a | Inclusion date |
| 906a | Governing source |
| 690 5 | Harvard Library |
The last three fields are Harvard-specific (per Harvard's documentation) and can be easily excluded. Additional fields can also be easily added. See the Library of Congress' official documentation on MARC21 for standard field codes.
This script expects source files will be in data/mrc/ and will output the converted files to data/csv/. Simply run:
python convert.py
The script requries pymarc. Install it with:
pip install pymarc