[Primo] - frbr - improve algorythm - consider all keys and not only the first matching key
In Primo FRBR documentation is explained that (I added ** signs):
In the FRBRization process, the keys are compared. If a record has a matching key with another record, it is added to the same FRBR group. **Once a match is found, the system does not continue searching** for matches since a record can belong to one FRBR group only.
In our experience this may lead to false positives. When comparing keys between frbr candidates, the system should consider instead the whole set of matching keys, not just the first matching one. And for example do not group records having more non matching keys than matching keys.
In attachment an example where the books share the title and only 1 of the authors. This is not uncommon in case of exhibitions catalogues where the artist works together with the curators. FRBR considers the 2 different books as 2 different versions.
Here are matching keys of the record proposed as example:
Record 1 (2006):
<frbr><t>1</t><k1>$$Kubzroyoux jean christophe$$AA</k1><k1>$$Kubzdean tacita$$AA</k1><k1>$$Kubzwarner marina$$AA</k1><k1>$$Kubzgreer germaine$$AA</k1><k3>$$Kubztacita dean$$AT</k3></frbr>
<frbr><t>1</t><k1>$$Kubzde cecco emanuela$$AA</k1><k1>$$Kubzdean tacita$$AA</k1><k3>$$Kubztacita dean$$AT</k3></frbr>
Luigi Siciliano commented
Let me add that the documentation section I quoted in this idea, explaining the current behavior, is the chapter "The FRBR Vector" in the Technical Guide.