Inconsistencies in update queries

[joachimvh]

I have found some inconsistencies in the output of some update queries.

1. Some updates have a type, while others have an updateType instead.
2. In case of the default graph (or any other reference to graphs that is not an IRI), sometimes it has type graph, and sometimes it doesn't.
3. This one is more debatable, but graphs are represented differently than in SELECT queries. For SELECT queries, you have a graph statement with a nested BGP statement. For updates, the graph statement immediately contains the triples. In the first case the graph can contain any possible SPARQL statement and in the second case only triple patterns though, so I can see why this one might be done this way.