Upload triples into Fuseki dataset avoiding IRI clashes

Question

I have created a dataset using the web GUI of Apache Jena Fuseki and now I would like to programmatically upload some triples into the remote dataset using RDFConnection .

The triples that I would like to upload come from different sources and I want to avoid creating RDF nodes that have IRIs already present in the remote dataset. Is there a way to do that without explicitly querying the remote dataset? Returning a list of all the IRIs is a waste of resources.

For instance, if I have this dataset (in Turtle syntax) in my remote triple store:

<iri_1> a foaf:Person
<iri_2> a foaf:Person
...
<iri_n> a foaf:Person

and I would like to add a new individual of foaf:Person which is not the same of the existing ones, how do I generate an IRI for the new individual without the risk of using one already existing in the remote dataset?

Answer 1

From your example it looks like each subject is named after the number of occurrences of the class. To keep doing this, I would count the number of foaf:Persons there are (trivial and quick sparql query) and just offset new subjects from the count.

If you can't contact the database then I would just generate uuids, which should never clash. The subjects won't be visually appealing, but they certainly won't clash. You should be able to use UUID out of java.util . RDFUnit has an example of generating unique IRIs using JennaUUID , which should also work for you.