Presence of Empty Containers

[eliarizzetto]

We have detected the presence of empty containers, i.e. resources of type fabio:JournalVolume or fabio:JournalIssue that are not linked to a journal article by an incoming frbr:partOf relation.

2 likely scenarios have been hypothesised to explain these cases:

1. The container is mentioned as a citing/cited entity in a citation provided by one of the authoritative primary sources, therefore it is assumed to be correct and the citations should be available in OpenCitations Index. In this case, we expect the container to have an external ID (e.g. DOI), and we would maintain it in Meta despite it being empty.
2. The emptiness of the container results from the an error in the deduplication process. According to this hypothesis, for duplicate journal articles that have been merged together and appeared to be contained in different (yet duplicate) journal volumes/issues, the system succeeded in the correct deduplication of the articles but failed in deduplicating their containers, which therefore would have remained empty. In this case, all the chain of entities making up the containment relationship might possibly need a merge, i.e. if a journal volume does not contain any article, and the issue containing it does not contain any other volume or article, then both the issue and the volume in question might need to be merged with their respective duplicate.

N.B. Cases of empty containers might include other types of container resources (which have not been checked yet): Journal, BookSeries, AcademicProceedings, Book, ReferenceBook, Series, and the superclass Expression.

Examples of empty journal volumes and issues can be retrieved with the following SPARQL query:

PREFIX frbr: <http://purl.org/vocab/frbr/core#>
PREFIX fabio: <http://purl.org/spar/fabio/>
PREFIX datacite: <http://purl.org/spar/datacite/>

SELECT ?container {
  ?container a ?type.
  VALUES ?type { fabio:JournalIssue fabio:JournalVolume}
  
  FILTER NOT EXISTS {
    ?x frbr:partOf ?container;
      a fabio:Expression;
      datacite:hasIdentifier ?identifier.
  }
}
LIMIT 10