Bulk replacing statements in ontotext GraphDB

Question

I am currently working on integrating with ontotext's GraphDB using their RDF4J API.

I am trying to insert and replace large sets of statements, ideally in one go. I am currently doing this on my Fuseki instance (which I'm migrating away from) by creating this block for every triple I want to insert into my Graph (currently only using default graph) and submitting that as one update call using semi-colons to separate the statements (where that is going into the post body);

delete { <x:test> <y:name> ?o } 
insert { <x:test> <y:name> "Test" } 
where { optional{ <x:test> <y:name> ?o }};
delete { <x:test> <y:description> ?o } 
insert { <x:test> <y:description> "Test" } 
where { optional{ <x:test> <y:description> ?o }};
delete { <x:test2> <y:name> ?o } 
insert { <x:test2> <y:name> "Test" } 
where { optional{ <x:test2> <y:name> ?o }};

I have been looking at the API and as far as I can work out Statement Post Method (POST /repositories/{repositoryID}/statement) only allows a SPARQL 1.1 Update string like that in the query parameter called update.

This means I would end up appending thousands of these blocks of SPARQL onto the query parameter which really doesn't feel right.

Am I missing something with how you update statements in GraphDB? Is there a better strategy I should be following?

EDIT 1

After a play around I've created the following which seems to work however I can't say whether or not it would be considered a good way of doing it;

DELETE {
   <x:test> ?p ?o
}
WHERE {
    <x:test> ?p ?o . 
        VALUES (?p) {
          (<y:name>) 
          (<y:description>)
        } 
};
INSERT DATA {
    <x:test> <y:name> "New Name" .
    <x:test> <y:description> "New Description" .
}

Any help would be greatly appreciated.

Many Thanks, John

Answer 1

You can sequence multiple SPARQL updates together using a semicolon, ";". For example:

delete { <x:test1> <y:name> ?o } 
insert { <x:test1> <y:name> "Test 1" } 
where { <x:test1> <y:name> ?o }};
delete { <x:test2> <y:name> ?o } 
insert { <x:test2> <y:name> "Test 2" } 
where { <x:test2> <y:name> ?o }

Just send such a sequence string to the update endpoint and it will be executed in one go.

As for whether this is the most efficient way to do it, that's a different matter. Do the statements that you are trying to replace share any characteristics? If so, it might be possible to express the update in a single query, rather than a sequence for each triple / subject.