RDF Views (Virtual Knowledge Graphs) in Virtuoso

By default, Virtuoso stores all RDF data in a single "RDF Storage" that consists of one four-column relational table containing all triples of all graphs. System Administrators can extend this basic schema with additional Quad Mapping Storage Types to represent a variety of SQL tables and views stored in Virtuoso or in attached data sources.

Understanding Quad Mapping

Virtuoso uses a layered architecture for mapping relational data to RDF:

  • Quad Mapping Storage Type: Container for quad mappings (e.g., DefaultQuadStorage, DefaultServiceStorage)
  • Quad Map Collection: Configuration that defines how relational data is mapped to RDF
  • Knowledge Graph: The destination where resulting triples are stored
  • Data Sources: The relational tables or SQL queries that provide the source data
DefaultQuadMap
DefaultQuadStorage
DefaultServiceStorage
SyncToQuads
TmpQuadStorage

Quad Mapping Storage Type: DefaultQuadMap

This pre-defined storage contains tables to quad maps collectively known as "RDF Views" or "Virtual Knowledge Graphs". It contains 65 top-level quad maps.

Oracle HR
SQL Server Northwind
Informix Stores

Knowledge Graph: oracle-hr

Quad Map: Countries RDF View

Data sources: Multiple Tables

Click to show SQL data sources ▼

  • "hr"."ora18"."COUNTRIES"
  • "hr"."ora18"."LOCATIONS"
Quad Map: Departments RDF View

Data sources: Multiple Tables

Click to show SQL data sources ▼

  • "hr"."ora18"."DEPARTMENTS"
  • "hr"."ora18"."LOCATIONS"
  • "hr"."ora18"."EMPLOYEES"
Quad Map: Employees RDF View

Data sources: Multiple Tables

Click to show SQL data sources ▼

  • "hr"."ora18"."EMPLOYEES"
  • "hr"."ora18"."JOBS"
  • "hr"."ora18"."DEPARTMENTS"

Knowledge Graph: northwind-sqlserver

Quad Map: Orders RDF View

Data sources: Multiple Tables

Click to show SQL data sources ▼

  • "sqlserver"."northwind"."orders"
  • "sqlserver"."northwind"."customers"
  • "sqlserver"."northwind"."employees"
  • "sqlserver"."northwind"."order_details"
Quad Map: Products RDF View

Data sources: Multiple Tables

Click to show SQL data sources ▼

  • "sqlserver"."northwind"."products"
  • "sqlserver"."northwind"."suppliers"
  • "sqlserver"."northwind"."categories"
  • "sqlserver"."northwind"."order_details"
Quad Map: Region RDF View

Data sources: Multiple Tables

Click to show SQL data sources ▼

  • "sqlserver"."northwind"."region"
  • "sqlserver"."northwind"."territories"

Knowledge Graph: informix-demo

Quad Map: Customer Calls RDF View

Data sources: Single Table

Click to show SQL data sources ▼

  • "informix"."stores"."cust_calls"
Quad Map: Customer RDF View

Data sources: Single Table

Click to show SQL data sources ▼

  • "informix"."stores"."customer"
Quad Map: VoidStatistics RDF View

Data sources: Multiple Tables

Click to show SQL data sources ▼

  • "informix"."stores"."cust_callsCount"
  • "informix"."stores"."customerCount"
  • "informix"."stores"."itemsCount"
  • "informix"."stores"."manufactCount"
  • "informix"."stores"."informix-demo__Total"
  • "informix"."stores"."employeeCount"

Quad Mapping Storage Type: DefaultServiceStorage

This pre-defined storage is used internally for SERVICE {...} clauses of SPARQL queries; not for regular use.

The storage contains no RDF Views.

Default Quad Map Storage Type: DefaultServiceMap

This Quad Map provides triples for various graphs named by format default-iid-nonblank

The data come from system procedure views that compose and send HTTP requests to remote SPARQL service endpoints and then parse answers.

Quad Mapping Storage Type: SyncToQuads

This pre-defined storage is to enumerate quad maps mentioned by triggers based on RDF Views; these triggers track changes in source tables of RDF Views and materialize updated content of views as "physical" triples.

It contains 12 top-level quad maps ("RDF Views") grouped by target Knowledge Graph:

Knowledge Graph: csv_northwind ▼
Quad Map Data Source
Categories "csv"."northwind"."categories"
Customers "csv"."northwind"."customers"
Products "csv"."northwind"."products"

The storage has no default quad map.

Quad Mapping Storage Type: TmpQuadStorage

This storage is created by some application or by system administrator.

The storage contains no RDF Views.

Default Quad Map Storage Type: TmpQuadMap

This Quad Map provides triples for various graphs named by format default-iid-blank

The data come from DB.DBA.RDF_QUAD_TMP

Quad Mapping Storage Type: DefaultQuadStorage

This Quad Map Storage Type contains physical triples. It is used for storing RDF data loaded from files or created by SPARQL/UPDATE operations.

About DefaultQuadStorage

This storage type is responsible for physical triples only, unlike DefaultQuadMap which handles virtual triples from quad mappings.