Great, thanks very much Stephen for the comprehensive answer.
Bill On 24 Aug 2012, at 20:52, Stephen Allen <[email protected]> wrote: > On Fri, Aug 24, 2012 at 11:42 AM, Bill Roberts <[email protected]> wrote: >> With Fuseki, if I send a SPARQL UPDATE to the 'update' endpoint that >> contains more than one operation, eg >> >> >> >> DELETE {…pattern…} WHERE {…} ; >> INSERT DATA { …some triples…} >> >> what is the scope of the transaction? Does it include both of the >> operations? Or does each operation happen in a separate transaction? >> >> (Note that I am thinking of a case that can't easily be merged into a single >> DELETE..INSERT..WHERE pattern so does require more than one operation} >> > > The relevent part of SPARQL 1.1 Update [1] is: > Each request SHOULD be treated atomically by a SPARQL 1.1 > Update service, i.e., the execution of all the operations in one > request should be performed atomically. Any resulting concurrency > issues will be a matter for each implementation to consider > according to its own architecture. > > Fuseki will perform multiple update operations in a single request > atomically. It depends on the underlying store as to whether or not > this is a full ACID transaction. TDB will use transactions. > > -Stephen > > [1] http://www.w3.org/TR/sparql11-update/#updateServices
