Truncated blank node identifiers?

[marijane]

While mapping various datasets to the VIVO-ISF ontology, I am noticing that some blank nodes are not getting the long generated blank node identifiers, rather they only have the name prefix, which is having the effect of collapsing things into one blank node. Here is example RDF from mapping a SciENcv profile to VIVO-ISF, where blank nodes of type vcard:Individual do not get unique identifiers:

_:vcard_Individual1 <http://www.w3.org/2006/vcard/ns#hasName> _:vcard_Name1_37619fc13053f82b7cb7da3d24ceb1598ab6d05c_N481_ff568e57abbe6ec3237143271826467f343bfe8b_N479 . _:vcard_Name1_37619fc13053f82b7cb7da3d24ceb1598ab6d05c_N481_ff568e57abbe6ec3237143271826467f343bfe8b_N479 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2006/vcard/ns#Name> . <https://www.ncbi.nlm.nih.gov/cv/marijanewhite> <http://purl.obolibrary.org/obo/ARG_2000028> _:vcard_Individual1 . <https://www.ncbi.nlm.nih.gov/cv/marijanewhite> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://xmlns.com/foaf/0.1/Person> . _:vcard_Individual1 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2006/vcard/ns#Individual> . _:vcard_Name1_37619fc13053f82b7cb7da3d24ceb1598ab6d05c_N481_ff568e57abbe6ec3237143271826467f343bfe8b_N479 <http://www.w3.org/2006/vcard/ns#givenName> "Marijane" . _:vcard_Name1_37619fc13053f82b7cb7da3d24ceb1598ab6d05c_N481_ff568e57abbe6ec3237143271826467f343bfe8b_N479 <http://www.w3.org/2006/vcard/ns#familyName> "White" .

Note that the vcard:Name node has a proper identifier.

You can see the R2RML file at https://raw.githubusercontent.com/openrif/sciencv-integration/master/code/scv2vivo/vcard.ttl
Karma report: https://github.com/openrif/sciencv-integration/blob/master/code/scv2vivo/vcard.md
Printed model: https://github.com/openrif/sciencv-integration/blob/master/code/scv2vivo/vcard.png

I have another dataset where this happened both with vcard:Identifier and vivo:Position. A pattern I think I might be seeing here is that blank nodes that are one hop away from nodes with URIs are not getting generated IDs. Blank nodes that are more than one hop away, with another blank node between them and the node with the URI, seem to be getting the expected identifiers. This could be a red herring, though.

Environment:
OS X 10.11.6 (15G1421)
Google Chrome Version 57.0.2987.133 (64-bit)
Karma version 2.052

[dkapoor]

Sorry about the delay in responding to this. The TTL file does seem correct and generates Individual as a Blank Node.
Can you attack a small sample input file and your karma home folder so we can reproduce this at our end? I could not find those in your github repo.

Thanks.

[megankatsumi]

I am having a similar issue regarding the generation of blank nodes. For example, I have a dataset where I would like one of the columns to be related to a blank node, however when I create the mapping it appears as though Karma generates a single blank node for all of the rows. My goal is to map the ID field to an individual of type "NodePD", which is related to some anonymous individual of type "Node" via the "hasManifestation" property, and so on.
Setting aside the question of whether this is a good design, I am wondering how I can modify the mapping so that for each row, the NodePD individual will be related to a different, anonymous individual of type Node. Currently, the mapping generates a single _:Node1 blank node for all of the rows in the data.

I have attached a screenshot along with the R2RML and RDF output files for your reference.
Archive.zip

[marijane]

I also need to apologize for my delayed follow up here, I have been extremely busy!

You can see the missing identifier on _:vcard_Individual1 in the snipped I shared. It doesn't have a suffix, and so all the Individual nodes get collapsed into one. Compare this with _:vcard_Name1_37619fc13053f82b7cb7da3d24ceb1598ab6d05c_N481_ff568e57abbe6ec3237143271826467f343bfe8b_N479 which does have the suffix.

I unfortunately cannot share my input files because they contain PII for people other than myself. If I get a chance I will try to create a test set and reproduce.

[megankatsumi]

Has there been any progress on this issue? I am encountering the same thing. Sometimes Karma generates a "full" blank node whereas other times the blank node appears to be truncated, making a mess of the resulting RDF. I'm not sure what the cause is, I can only assume it has something to do with certain types of models.