A few weeks ago, I shared an excited tweet with the archives twitterverse announcing that I had successfully tested importing a MARC record from Archivists’ Toolkit into WorldCat. The tweet garnered more attention than I had anticipated, including a few direct messages from fellow archivists wanting to know how we came up with a solution to the MARC from AT problem. Here is what we did.
The problems with MARCXML exported from AT are few but significant. My colleague Mark Custer at ECU recently posted to the AT user group listserv a question about the fact that AT does not currently allow subfields for subject headings, so the MARC from AT is missing the subfield indicators. I set up a meeting with a cataloger at my library to help me look at the MARCXML files being exported from AT to see what her thoughts were about whether the records could be considered complete. We took a look at MARC for archival material already on WorldCat and compared that to what we exported from AT. She illustrated what she saw as the issues that would prevent proper sharing of the MARC with our local catalog and WorldCat:
- Missing fixed fields including Ctrl, Desc, and Date (if no date range was included in the finding aid)
- Missing subject heading subfield delimiters
- 650 used instead of 600 field in some instances
- Missing indicators for 245 (and 545, optional)
- Missing cataloging source for 049 and 040
Because the MARC exported from AT is in MARCXML format and our catalogers work with the MRC format, we used MARCedit to convert the record from MARCXML to MRC. Once these missing and erroneous elements were fixed using MARCedit, we were ready to test import the record. Our library’s account with OCLC Connexion accepts imported records in DAT format, so we saved the MRC file as a DAT file. We tried uploading to Connexion using local bibliographic import and were successful. We determined that it would probably be easier to edit the MARC directly in Connexion, so we will do that in the future. The cataloger and I decided to upload the file to WorldCat as an official record, which worked, as well as to our local catalog, which also worked!
One issue for my library is that our finding aids are missing subject terms and authority work that most catalogers would require for submission to WorldCat. We have started incorporating this cataloger into our processing workflow and introduced her to the Names and Subjects modules in AT so that she can finalize subject headings and names that we assign. We can also consider an automated batch update for all our exported MARCXML to include the edits listed above, incorporating help from our technology team and their knowledge of FTP and scripting. In the meantime, we will be submitting our MARC one at a time since our finding aids are incomplete.
Here’s a recap of our tentative workflow, for your information:
- Open MARCedit, then Tools
- Choose MARCXML file as input file
- Tell program output file name (copy and paste input file info; change ending to .mrc)
- Select MARC21XML to MARC plus Translate to MARC8
- Select Execute
- Open OCLC Connexion
- Import records; browse to .mrc file
- Edit directly in OCLC Connexion
- Update fixed fields including Ctrl, Desc, and Date
- Change 650 to 600 when necessary
- Add subfield delimiters to subject headings
- Add indicators to 545, 245 as needed
- Add cataloging source to 040 and 049
- Save and validate
- Login to OCLC, select ActionHoldingsUpdateHolding to load directly to WorldCat
Thoughts, comments, ideas, and suggestions are gratefully welcomed! I am really curious to know how others approach this issue.