search

lcnetdev / marc2bibframe2

Convert MARC records to BIBFRAME2 RDF
http://www.loc.gov/bibframe/
Creative Commons Zero v1.0 Universal
88 stars 35 forks source link

Apparent duplication of field rdf:type #218

Closed laura-ake closed 2 years ago

laura-ake commented 2 years ago

Can you please check the code that generates this section of the converted record? It appears a line is being duplicated, and may be out of place: Tidsskrigt for Den norske legeforening MARCXML 2022_01_12.txt

</bf:Work>

online resource It also appears to my eyes that in addition to the duplicated elements, there's another element referring to the same bf:instance with a different value. This comes from conversion of an electronic journal record (so far, we haven't found an example with a print journal). I have attached the marcxml record from our Ex Libris Alma catalog and a copy of the latest converted output from Alma (which incorporates marc2bibframe2 in its tools). We confirmed via a separate installation of marc2bibframe2 that the duplication is happening for this record outside Alma. [Tidsskrift for Den norske legeforeningThe Journal of the Norwegian Medical Association.bibframexml_2022_01_12.txt](https://github.com/lcnetdev/marc2bibframe2/files/7856457/Tidsskrift.for.Den.norske.legeforeningThe.Journal.of.the.Norwegian.Medical.Association.bibframexml_2022_01_12.txt) Thank you. Laura Akerman for the IGELU-ELUNA Linked Open Data Community of Practice Working Group
wafschneider commented 2 years ago

@laura-ake thank you for this feedback, I apologize for the long delay in getting back to you.

The first rdf:type property on the bf:Instance comes from the leader. The MARC to BIBFRAME specifications state that if LDR/07 is "s" and LDR/06 is "a", a triple of Instance - rdf:type - bf:Print should be created.

The second and third rdf:type properties are coming from the two 007 fields in the record:

<controlfield tag="007">cr-bn||||||abp</controlfield>
<controlfield tag="007">cr-bn||||||ada</controlfield>

It looks like this is an error in the converter, probably left over from a previous version of the specifications, as the current specs do not have this direction, instead specifying that bf:media and bf:carrier properties be created if they are not already created by a 337 or 338. The implementation looks incorrect to me.

@jodiw01 -- if you can confirm my understanding I can fix this issue. Sorry for the inconvenience!

laura-ake commented 2 years ago

Thanks Wayne. Let me know if I can help with testing a fix, if I can remember what our example was when that happens.

Laura