[SOLVED] Amazon Neptune overwrites blank node values when inserting data

Amazon Neptune overwrites blank node values when inserting data

When inserting a triple with a blank node into an Amazon Neptune instance:

INSERT DATA { <http://example.com/s> <http://example.com/p> _:something }

…the blank node's value gets replaced with b<SOME NUMBER>. Running a query like this:

SELECT ?o WHERE { <http://example.com/s> <http://example.com/p> ?o }

…will return something like this:

{
    "results": {
        "bindings": [
            {
                "o": {
                    "type": "bnode",
                    "value": "b24508943"
                }
            }
        ]
    }
}

Is there a way to retain original blank node names when loading data into Neptune?

Solution

The idea of blank nodes is that they are not addressable (from "outside"). If you need something you can address (i.e., "point to") later, you should give it a URI.

If you have a blank node that serves as the subject of some statements, you might be able to find those statements if some other statement uses that blank node as well (e.g., as its object). That is, you could use a query to find them. Otherwise, use a URI.