Thank you for your comment.

> In Section 3.1.4 LOAD:
> where it says: "The LOAD operation reads an RDF document from a IRI and
> inserts its triples" and a little later where it says "the resulting
> triples will be inserted" I suggest s/insert/merge/ to be clear that
> existing triples are deleted when the new triples are LOADed.  
> Thus those phrases would read: "The LOAD operation reads an RDF document
> from a IRI and merges its triples"  and "the resulting triples will be
> merged".

Since both "insert" and "merge" imply new data and do not suggest the removal of information, then am I correct to presume that when you said, "existing triples are deleted," that you actually intended to say, "existing triples are not deleted"?

When the LOAD operation is executed, new data is added to the existing triples in a graph. No triples will be deleted from the destination graph. We believe that the word "insert" in the informal prose implies this. The use of the word "merge" would imply "RDF Merge". This may be essentially what a LOAD does, however the operation is described completely independently of RDF Merge. The formal definition of LOAD is in section 4.3.4, but to make it clearer in the prose we have added the following:

"If the destination graph already exists, then no data in that graph will be removed."

We would be grateful if you would acknowledge that your comments have been answered by sending a reply to this mailing list.

Regards, Paul Gearon, on behalf of the SPARQL WG.

Last modified on 1 August 2012, at 14:46