com.bigdata.rdf.sparql.ast

Interface StatementPatternNode.Annotations

All Superinterfaces:

GroupMemberNodeBase.Annotations, IJoinNode.Annotations, IQueryNode.Annotations

All Known Subinterfaces:

AbstractOptimizerTestCase.Annotations

Enclosing class:

StatementPatternNode

public static interface StatementPatternNode.Annotations
extends GroupMemberNodeBase.Annotations, IJoinNode.Annotations

Field Summary

Fields

Modifier and Type	Field and Description
static boolean	DEFAULT_DISTINCT
static String	DISTINCT Boolean flag indicates that the distinct solutions for the statement pattern are required (false).
static String	DISTINCT_TERM_SCAN_VAR An optional annotation whose value the variable that will be bound by a DistinctTermAdvancer layered over the access path.
static String	EXISTS The existence of at least one solution will be verified otherwise the solution will be failed.
static String	FAST_RANGE_COUNT_VAR An optional annotation whose value is a variable which will become bound to the fast range count of the associated triple pattern.
static String	INLINE The data for this access path is inline.
static String	RANGE An optional attribute whose value is an RangeBOp which models the key-range constraint on the access path.
static String	SCOPE The StatementPattern.Scope (required).
static String	SID An optional annotation whose value is the variable which will be bound to the statement identifier for the matched statement patterns.

Fields inherited from interface com.bigdata.rdf.sparql.ast.IQueryNode.Annotations

QUERY_HINTS

Fields inherited from interface com.bigdata.rdf.sparql.ast.IJoinNode.Annotations

DEFAULT_MINUS, DEFAULT_OPTIONAL, FILTERS, MINUS, OPTIONAL

Field Detail

SCOPE

static final String SCOPE

The StatementPattern.Scope (required).

See Also:

ASTGraphGroupOptimizer, Constant Field Values

DISTINCT

static final String DISTINCT

Boolean flag indicates that the distinct solutions for the statement pattern are required (false).

Note: This is a hint that the DistinctTermAdvancer should be used to visit the distinct ISPOs having a common prefix. This is used for GRAPH ?g {}, which evaluates to all of the named graphs in the database (if the named graphs were not explicitly specified).

Note: For only partly historical reasons, this is not used to mark default graph access. A default graph access path strips the context and then applies a DISTINCT filter to the resulting triples.

See Also:

Constant Field Values

DEFAULT_DISTINCT

static final boolean DEFAULT_DISTINCT

See Also:

Constant Field Values

EXISTS

static final String EXISTS

The existence of at least one solution will be verified otherwise the solution will be failed. This turns into an iterator with a limit of ONE (1) on the SPOAccessPath.

Note: This is used in combination with a join against an inline access path for the named graphs. The "exists" statement pattern MUST run after the access path which produces the variety since it will be used to constrain that as-bound variety. This the join order in query plan must look like:

 (_,_,_,?g)[@INLINE,@IN(g,namedGraphs)] x (_,_,_,?g)[@EXISTS]
 

rather than

 (_,_,_,?g)[@EXISTS] x (_,_,_,?g)[@INLINE,@IN(g,namedGraphs)]
 

as the latter will find only one solution for ?g.

See Also:

(Optimization for GRAPH uri {} and GRAPH ?foo {}), Constant Field Values

INLINE

static final String INLINE

The data for this access path is inline. The value of the attribute is the column projection / solution set reference. TODO A column projection is more efficient when we are handling things like the named graphs or constraining the subquery for an optional with multiple predicates. That column projection can be modeled as IN(var,values). The IN filter could be attached to IJoinNode.Annotations.FILTERS or it could be the value of this attribute.

We also have use cases for inline solution set access paths for use with the samples materialized by the RTO. Those should be an HTree and the data should be modeled as ISolutions. (Note that some vertices may correspond to "bop fragment" joins, in which case the can not be modeled as ISPOs.)

Both the column projection (IN) and the inline solution set (HTree) are simpler access paths. They only support element visitation, a full scan of the access path (this is the same as saying that there are no join variables), or probing to find all solutions which join on some join variable(s). This is in contrast to the SPOAccessPath, which supports key-range constraints (prefix) and range constraints (prefix with key range on a data type value).

See Also:

(Optimization for GRAPH uri {} and GRAPH ?foo {}), Constant Field Values

RANGE

static final String RANGE

An optional attribute whose value is an RangeBOp which models the key-range constraint on the access path. The RangeBOp is used when there are filters which impose a GT/GTE and/or LT/LTE restriction on the values which a variable may take on for that access path. TODO We should also handle datatype constraints on a variable here. For example, if a variable is known to be numeric, or known to be xsd:int, then we can immediately reject any bindings which would violate that type constraint. To do this right, we need to notice those type constraints and propagate them backwards in the plan so we can reject bindings as early as possible. (In fact, we can also do a range constraint which spans each of the datatypes which the variable could take on. Datatype constraints and value range constraints are very much related. The datatype constraint is effectively a value range constraint allowing the entire value space for that datatype. Likewise, a value range constraint must be applied across the UNION of the allowable ground datatypes for the variable.)

See Also:

ASTRangeConstraintOptimizer, (lift range constraints onto AP), Constant Field Values

SID

static final String SID

An optional annotation whose value is the variable which will be bound to the statement identifier for the matched statement patterns. The statement identifier is always formed from the subject, predicate and object (the triple). The context is NOT represented in the statement identifier. This keeps the semantics consistent with RDF reification.

See Also:

Reification Done Right, Constant Field Values

FAST_RANGE_COUNT_VAR

static final String FAST_RANGE_COUNT_VAR

An optional annotation whose value is a variable which will become bound to the fast range count of the associated triple pattern.

See Also:

SELECT COUNT(...) (DISTINCT|REDUCED) {single-triple-pattern} is slow. , Constant Field Values

DISTINCT_TERM_SCAN_VAR

static final String DISTINCT_TERM_SCAN_VAR

An optional annotation whose value the variable that will be bound by a DistinctTermAdvancer layered over the access path.

See Also:

DISTINCT PREDICATEs query is slow , Constant Field Values