Using SPARQL 1.1 INSERT patterns and CONSTRUCT to Enhance Existing Data

What?

Using SPARQL 1.1 to provide new insights into existing data.

Why?

There are times when existing data doesn't provide the required insights for a given audience or data analytics pursuit.

How?

Use SPARQL 1.1 INSERT patterns to generate new data from existing data. Basically, this is like make a CONSTRUCT query where the output is persisted to a specific Named Graph.

The steps are as follows:

1. Assume the following Raw Data Representation in Turtle:
   

   
   <#rojer>    <#contactType> "100 - Sport" .
   <#alice>    <#contactType> "98 - School" .
   <#bob>      <#contactType> "99 - University" .
   <#john>     <#contactType> "101 - Other" .
   <#carol>    <#contactType> "100 - Friends" .
   <#karine>   <#contactType> "102 - Colleagues" .
   <#ivan>     <#contactType> "99 - University" .
   <#gloria>   <#contactType> "101 - Other" .
   <#kate>     <#contactType> "101 - Other" .
   <#jordan>   <#contactType> "99 - University" .
   <#brigitte> <#contactType> "100 - Sport" .
   

   
   
2. Load the sample data using SPARQL 1.1. INSERT Syntax:
   

   
   INSERT 
     {
       GRAPH <urn:sparql:tests:people> 
         { 
           <#rojer>    <#contactType> "100 - Sport" .
           <#alice>    <#contactType> "98 - School" .
           <#bob>      <#contactType> "99 - University" .
           <#john>     <#contactType> "101 - Other" .
           <#carol>    <#contactType> "100 - Friends" .
           <#karine>   <#contactType> "102 - Colleagues" .
           <#ivan>     <#contactType> "99 - University" .
           <#gloria>   <#contactType> "101 - Other" .
           <#kate>     <#contactType> "101 - Other" .
           <#jordan>   <#contactType> "99 - University" .
           <#brigitte> <#contactType> "100 - Sport" .
         }
     }   
   

   
   
3. Let's get all Contact types:
   

   
   SELECT DISTINCT ?s2 as ?contact_type
   FROM <urn:sparql:tests:people> 
   WHERE 
     { 
       ?s1 <#contactType> ?s2 .    
     }   
   

   
   
4. View the SPARQL Query Definition via SPARQL Protocol URL;
5. View the SPARQL Query Results via SPARQL Protocol URL
6. CONSTRUCT Others: Purpose -- For every entity of the types from above to set to type: <#Friend> [a foaf:knows] when code does not start with '101 -':
   

   
   CONSTRUCT { ?s1 a <#Friend> } 
   FROM <urn:sparql:tests:people>
   WHERE 
    { 
      ?s1 <#contactType> ?s2 .
      FILTER ( !strStarts(?s2, "101 -") ). 
    }
    
    
   # Query result: 
   @prefix rdf:	<http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
   @prefix ns1:	<#> .
   ns1:rojer	rdf:type	ns1:Friend .
   ns1:karine	rdf:type	ns1:Friend .
   ns1:ivan	rdf:type	ns1:Friend .
   ns1:jordan	rdf:type	ns1:Friend .
   ns1:brigitte	rdf:type	ns1:Friend .
   ns1:alice	rdf:type	ns1:Friend .
   ns1:bob	        rdf:type	ns1:Friend .
   ns1:carol	rdf:type	ns1:Friend .
   

   
   
7. Generate Friends list:
   

   
   INSERT  
     { 
       GRAPH <urn:sparql:tests:friends> 
         { 
           ?s1 a <#Friend>  
         } 
     }
   FROM <urn:sparql:tests:people>  
   WHERE 
    { 
      ?s1 <#contactType> ?s2 .
      FILTER ( !strStarts(?s2, "101 -") ). 
    }
   

   
   
8. Let's get all Friends:
   

   
   SELECT * 
   FROM <urn:sparql:tests:friends>
   WHERE 
     { 
       ?s ?p ?o
     }
   

   
   
9. View the SPARQL Query Definition via SPARQL Protocol URL;
10. View the SPARQL Query Results via SPARQL Protocol URL

Related

• SPARQL 1.1. Features Examples Collection
• SPARQL 1.1. Property Paths
• SPARQL inline data via use of VALUES clause
• Virtuoso SPARQL 1.1 Syntax Tutorial
• SPARQL 1.1 Sub-queries
• SPARQL 1.1 Specification
• SPARQL Protocol (HTTP based Query Service)
• Virtuoso Documentation
• Virtuoso Tips and Tricks Collection