Mapping Update Requests to the Formal Model

\n\n

In this section we show how to map Update Requests in the SPARQL 1.1. Update Language to Update Operations over the Graph Store as defined earlier in this section. This mapping assumes that in all Update requests, any PREFIXes have been expanded. Moreover, we assume that WITH clauses have been replaced by wrapping both the QuadPatterns in subsequent DELETE and INSERT clauses, and likewise - in the absence of USING and USING NAMED clauses - the GroupGraphPattern in the subsequent WHERE clause, into GRAPH patterns.

\n

The mapping from requests to Update Operations is defined in terms of the recursive translation function Tr(GS,R) which takes the Graphstore GS - as before executing the request - and an update request R as input and exands it to an Update Operation call as shown in the following table. The COPY, MOVE, and ADD operations are not mentioned explicitly here, since they are understood as shortcuts.

\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n

Table 1: Mapping from Update Requests to Update Operations

Update request R\n	\n Tr(GS,R) =
\n R1\n ; R2\n \n	\n Tr(Tr(GS, R1\n ), R2\n )
\n INSERT DATA\n QuadData\n	\n OpInsertData(GS, QuadData)
\n DELETE DATA\n QuadData\n	\n OpDeleteData(GS, QuadData)
\n DELETE\n QuadPattern\n DEL\n INSERT\n QuadPattern\n INS\n \n UsingClause* \n WHERE\n GroupGraphPattern\n	\n OpDeleteInsert(GS, TrDataset\n (GS,UsingClause*), QuadPattern\n DEL, QuadPattern\n INS, GroupGraphPattern)
\n DELETE\n QuadPattern\n DEL\n \n UsingClause* \n WHERE\n GroupGraphPattern\n	\n OpDeleteInsert(GS, TrDataset\n (GS,UsingClause*), QuadPattern\n DEL, {}, GroupGraphPattern)
\n INSERT\n QuadPattern\n INS\n \n UsingClause* \n WHERE\n GroupGraphPattern\n	\n OpDeleteInsert(GS, TrDataset\n (GS,UsingClause*), {}, QuadPattern\n INS, GroupGraphPattern)
\n DELETE WHERE\n QuadPattern\n	\n OpDeleteInsert(GS, GS, QuadPattern, {}, QuadPattern)
\n LOAD (SILENT)?\n IRIref\n	\n OpLoad(GS, IRIref)
\n LOAD (SILENT)?\n IRIref\n from\n INTO GRAPH\n IRIref\n to\n	\n OpLoad(GS, IRIref\n from, IRIref\n to)
\n CLEAR (SILENT)? GRAPH\n IRIref\n	\n OpClear(GS, IRIref)
\n CLEAR (SILENT)? DEFAULT\n	\n OpClear\n def(GS)
\n CLEAR (SILENT)? NAMED\n	\n OpClear\n named(GS)
\n CLEAR (SILENT)? ALL\n	\n OpClear\n all(GS)
\n CREATE (SILENT)? GRAPH\n IRIref\n	\n OpCreate(GS, IRIref)
\n DROP (SILENT)? GRAPH\n IRIref\n	\n OpDrop(GS, IRIref)
\n DROP (SILENT)? DEFAULT\n	\n OpDrop\n def(GS)
\n DROP (SILENT)? NAMED\n	\n OpDrop\n named(GS)
\n DROP (SILENT)? ALL\n	\n OpDrop\n all(GS)

\n\n

This table uses one auxiliary translation function TrDataset() which constructs a dataset from the optional set of USING and USING NAMED clauses and is defined as follows:

\n\n \n \n \n \n \n \n \n \n \n

Table 2: Mapping \n UsingClause\n s to RDF Datasets

Translation Function	Definition
\n TrDataset\n (GS,\n UsingClause\n *) =	\n \n the RDF Dataset DS described by the UsingClauses, if non-empty \n the RDF Dataset corresponding to the current state of GS, otherwise \n \n

\n\n

Note:

\n

How exactly an RDF Dataset is obtained from the

\n

USING and

\n

USING NAMED clauses (e.g. by dereferencing graph name IRIs and trying to retrieve them, or by picking those graphs from the existing Graph Store) is implementation dependent. Particularly, this specification does not mandate any assumptions about blank node identity beyond the consideration for the analogous

\n

FROM and

\n

FROM NAMED clauses in Section

\n

Specifying RDF Datasets of the SPARQL 1.1 Query Language specification.