Added ConstructClause to Statement struct

[apr94]

I want to make sure that the addition I am making to the Statement struct looks good.

[xllora]

This looks just right on track. I just left a couple of comments. To make sure I read it correctly.

[xllora]

Could you elaborate why you need a new working clause? The where clause of a construct is the same as the one on a regular select. I was wondering if it is just cleaner to just add the actual construct information containing the info to create the new triples. Or am I just mislead by the name?

[apr94]

So constructClauses captures only the declarations within a CONSTRUCT statement that specifies how the triples must be built. For instance, consider the following:

construct {?s "new_predicate"@[] ?o} into ?a from ?b where {?s "old_predicate"@[] ?o}

constructClauses simply captures the ?s "new_predicate"@[] ?o part. workingClauses will capture the ?s "old_predicate"@[] ?o as before. constructClauses will capture multiple such declarations (separated by the . operator) and constructClause also has support for multiple reification clauses. So in a way, constructClause is similar to data within the Statement struct (except it has to support literals, bindings and reification). Does this seem like the right thing to do?

[xllora]

Despite the name this is just the info on the construct right? Just want to make sure I got it right :)

[apr94]

Yep, it is just the information required to insert a new triple into the graph. For instance (a contrived example):

construct {?s ?p ?o; ?p1 ?o1} into ?a from ?b where {?s ?p ?o. ?s ?p1 ?o2}

will lead to the creation of the following instance of constructClauses:

[]*ConstructClauses{
  {
    SBinding: ?s,
    PBinding: ?p,
    OBinding: ?o,
    ReificationClauses: []*ReificationClause{
      {
         PBinding: ?p1
         OBinding: ?o1
      },
   },
}

I guess the name is confusing. Do you have any suggestions?