Semtech - Sparql

The SPARQL Query Language
CapGemini Tutorial
Martin Giese
24 April 2020
Martin Giese SPARQL, CapGemini 24 April 2020 1 / 17

SPARQL
▶ SPARQL Protocol And RDF Query Language

SPARQL

▶ Documentation:https://www.w3.org/TR/sparql11-overview/

SPARQL

▶ Query Language

SPARQL

▶ Query Language
▶ Update Language

SPARQL

▶ Query Language
▶ Update Language
▶ Protocol to submit queries to a server (“endpoint”)

SPARQL

▶ Query Language
▶ Update Language
▶ Formatting of query results

SPARQL

▶ Query Language
▶ Update Language
▶ …

SPARQL

▶ Query Language
▶ Update Language
▶ …
▶ Try it out:

SPARQL

▶ Query Language
▶ Update Language
▶ …
▶ Try it out:
https://www.w3.org/wiki/SparqlEndpoints

SPARQL

▶ Query Language
▶ Update Language
▶ …
▶ Try it out:
DBpedia http://dbpedia.org/sparql

SPARQL

▶ Query Language
▶ Update Language
▶ …
▶ Try it out:
Wikidata https://query.wikidata.org/

SPARQL

▶ Query Language
▶ Update Language
▶ …
▶ Try it out:
Musicbrainz http://dbtune.org/musicbrainz/snorql/

SPARQL

▶ Query Language
▶ Update Language
▶ …
▶ Try it out:
Musicbrainz http://dbtune.org/musicbrainz/snorql/
World DB http://sws.ifi.uio.no/d2rq/snorql/

Simple Examples
▶ DBpedia has information about actors, movies, etc. https://dbpedia.org/

▶ Web interface for SPARQL writing: http://dbpedia.org/sparql

Simple Examples

People called “Johnny Depp”

PREFIX foaf: <http://xmlns.com/foaf/0.1/>
SELECT ?jd WHERE {
?jd foaf:name "Johnny Depp"@en .
}

Simple Examples

People called “Johnny Depp”

SELECT ?jd WHERE {
}
Answer:
?jd
<http://dbpedia.org/resource/Johnny_Depp>

Simple Examples (cont.)
Films starring people called “Johnny Depp”

PREFIX dbo: <http://dbpedia.org/ontology/>
SELECT ?m WHERE {
?m dbo:starring ?jd .
}

Films starring people called “Johnny Depp”

PREFIX dbo: <http://dbpedia.org/ontology/>
SELECT ?m WHERE {
}
Answer:
?m
<http://dbpedia.org/resource/Dead_Man>
<http://dbpedia.org/resource/Edward_Scissorhands>
<http://dbpedia.org/resource/Arizona_Dream>
…

Titles of films by people called “Johnny Depp”

SELECT ?title WHERE {
?m rdfs:label ?title .
}

Titles of films by people called “Johnny Depp”

SELECT ?title WHERE {
?m rdfs:label ?title .
}
Answer:
?title
"Truposz"@pl
"Dead Man"@en
"El sueño de Arizona"@es
"Arizona Dream"@en
…
Names of people who co-starred with “Johnny Depp”
SELECT DISTINCT ?collab WHERE {
?m dbo:starring ?other .
?other foaf:name ?collab .
}

?m dbo:starring ?other .
?other foaf:name ?collab .
}
Answer:
?collab
"Al Pacino"@en
"Antonio Banderas"@en
"Johnny Depp"@en
"Marlon Brando"@en
…
Graph Patterns
The previous SPARQL query as a graph:
foaf:name
?jd "Johnny Depp"@en
ng
ta rri
dbo:s
?m dbo
:st
arr
ing
foaf:name
?other ?collab

Graph Patterns
The previous SPARQL query as a graph:
foaf:name
?jd "Johnny Depp"@en
ng
ta rri
dbo:s
?m dbo
:st
arr
ing
foaf:name
?other ?collab
Pattern matching: assign values to variables to make this a sub-graph of the RDF graph!

Graph with blank nodes
Variables not SELECTed can equivalently be blank:
foaf:name
"Johnny Depp"@en
ring
star
dbo:
dbo:
s tarr
i ng
foaf:name
?collab

Graph with blank nodes
Variables not SELECTed can equivalently be blank:
foaf:name
"Johnny Depp"@en
ring
star
dbo:
dbo:
s tarr
i ng
foaf:name
?collab
Pattern matching: assign values to variables and blank nodes to make this a sub-graph of the
RDF graph!

SPARQL Query with blank nodes
_:jd foaf:name "Johnny Depp"@en .
_:m dbo:starring _:jd .
_:m dbo:starring _:other .
_:other foaf:name ?collab .
}

}
The same with blank node syntax

_:m dbo:starring [foaf:name "Johnny Depp"@en] .
_:other foaf:name ?collab.
}
}

_:m dbo:starring [foaf:name "Johnny Depp"@en] .
_:m dbo:starring [foaf:name ?collab] .
}

}

[ dbo:starring [foaf:name "Johnny Depp"@en] ;
dbo:starring [foaf:name ?collab]
]
}
}

[ dbo:starring [foaf:name "Johnny Depp"@en] ,
[foaf:name ?collab]
]
}
Filters
▶ Groups may include constraints or filters

▶ Reduces matches of surrounding group where filter applies
▶ E.g.
{
?x a dbo:Place ;
dbp:population ?pop .
FILTER (?pop > 1000000)
}

Filters
▶ Groups may include constraints or filters

▶ Reduces matches of surrounding group where filter applies
▶ E.g.
{
?x a dbo:Place ;
dbp:population ?pop .
FILTER (?pop > 1000000)
}
▶ E.g.
{
?x a dbo:Document ;
dbp:abstract ?abs .
FILTER (lang(?abs) = no")
}

Optional Patterns
▶ Allows a match to leave some variables unbound (e.g. no data was available)
▶ A match is a partial function from variables to RDF terms

Optional Patterns
▶ Groups may include optional parts
▶ E.g.
{
?x a dbo:Document ;
dbp:date ?date .
OPTIONAL {
?x dbp:abstract ?abs .
}
}

Optional Patterns
▶ Groups may include optional parts
▶ E.g.
{
?x a dbo:Document ;
dbp:date ?date .
OPTIONAL {
?x dbp:abstract ?abs .
}
}
▶ ?x and ?date bound in every match, ?abs bound if there is a Norwegian abstract
▶ Groups can contain several optional parts, evaluated separately
Matching Alternatives
▶ A UNION pattern matches if any of some alternatives matches


▶ E.g.
{
{ ?book dc:creator ?author ;
dc:created ?date . }
UNION
{ ?book foaf:maker ?author . }
UNION
{ ?author foaf:made ?book . }
}


▶ E.g.
{
UNION
UNION
}
▶ Variables in matches union of variables in sub-patterns


▶ E.g.
{
UNION
UNION
}
▶ Variables in matches union of variables in sub-patterns
▶ Match of one pattern leaves rest of variables unbound

Four Types of Queries
SELECT Compute table of bindings for variables

SELECT ?a ?b WHERE {
[ dc:creator ?a ;
dc:creator ?b ]
}


[ dc:creator ?a ;
dc:creator ?b ]
}
CONSTRUCT Use bindings to construct a new RDF graph
CONSTRUCT {
?a foaf:knows ?b .
} WHERE {
[ dc:creator ?a ;
dc:creator ?b ]
}


[ dc:creator ?a ;
dc:creator ?b ]
}
CONSTRUCT {
?a foaf:knows ?b .
} WHERE {
[ dc:creator ?a ;
dc:creator ?b ]
}
ASK Answer (yes/no) whether there is ≥ 1 match


[ dc:creator ?a ;
dc:creator ?b ]
}
CONSTRUCT {
?a foaf:knows ?b .
} WHERE {
[ dc:creator ?a ;
dc:creator ?b ]
}
ASK Answer (yes/no) whether there is ≥ 1 match
DESCRIBE Answer available information about matching resources

Solution Modifiers
▶ Patterns generate an unordered collection of solutions

Solution Modifiers

▶ Each solution is a partial function from variables to RDF terms

Solution Modifiers

▶ SELECT treats solutions as a sequence (solution sequence)

Solution Modifiers

▶ Sequence modifiers can modify the solution sequence:

Solution Modifiers

▶ Order: …ORDER BY ?country DESC(?pop)

Solution Modifiers

▶ Projection: SELECT ?x ?y WHERE…

Solution Modifiers

▶ Distinct: SELECT DISTINCT ?x WHERE…

Solution Modifiers

▶ Reduce: SELECT REDUCE ?x WHERE…

Solution Modifiers

▶ Sub-sequence: …LIMIT 10 OFFSET 50

Solution Modifiers

▶ Sub-sequence: …LIMIT 10 OFFSET 50
▶ Applied in this order.

Some other features in SPARQL
Not discussed today:

▶ aggregate functions (count, sum, average,…)


▶ negation, set difference


▶ nested sub-queries


▶ federation: execute parts of queries on remote endpoints


▶ updates (add delete triples)


▶ updates (add delete triples)
Have a look at the SPARQL 1.1 docs…
https://www.w3.org/TR/sparql11-query/

SPARQL on the 'Net
▶ Many sites (dbpedia, wikidata, dbtunes,…) publish SPARQL endpoints

SPARQL on the 'Net

▶ I.e. SPARQL queries can be submitted to a database server that sends back the results

SPARQL on the 'Net

▶ RESTful HTTP requests to submit URL-encoded queries to server
GET /sparql/?query=... HTTP/1.1

SPARQL on the 'Net

▶ Server responds with an encoding of the result set, see
http://www.w3.org/TR/rdf-sparql-XMLres/

SPARQL on the 'Net

▶ Nothing you would want to do manually!

SPARQL on the 'Net

▶ Mostly, there is a web page for manual entry of queries

SPARQL on the 'Net

▶ Mostly, there is a web page for manual entry of queries
▶ RDF APIs like Jena or RDF4J can send queries to endpoints

Exercise: SPARQL
▶ Write SPARQL queries.

▶ Two possibilities:
▶ Run queries on SPARQL endpoint at http://sws.ifi.uio.no/d2rq/snorql/.
▶ Host your own triples with Fuseki
▶ Many queries, choose according to your skills.

Semtech - Sparql

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Semtech - Sparql

Uploaded by

Copyright:

Available Formats

The SPARQL Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 1 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ SPARQL Protocol And RDF Query Language

Martin Giese SPARQL, CapGemini 24 April 2020 2 / 17

▶ DBpedia has information about actors, movies, etc. https://dbpedia.org/

Martin Giese SPARQL, CapGemini 24 April 2020 3 / 17

▶ DBpedia has information about actors, movies, etc. https://dbpedia.org/

People called “Johnny Depp”

Martin Giese SPARQL, CapGemini 24 April 2020 3 / 17

▶ DBpedia has information about actors, movies, etc. https://dbpedia.org/

People called “Johnny Depp”

Martin Giese SPARQL, CapGemini 24 April 2020 3 / 17

Films starring people called “Johnny Depp”

Martin Giese SPARQL, CapGemini 24 April 2020 4 / 17

Films starring people called “Johnny Depp”

Martin Giese SPARQL, CapGemini 24 April 2020 4 / 17

Titles of films by people called “Johnny Depp”

Martin Giese SPARQL, CapGemini 24 April 2020 5 / 17

Titles of films by people called “Johnny Depp”

Martin Giese SPARQL, CapGemini 24 April 2020 6 / 17

The previous SPARQL query as a graph:

Martin Giese SPARQL, CapGemini 24 April 2020 7 / 17

The previous SPARQL query as a graph:

Martin Giese SPARQL, CapGemini 24 April 2020 7 / 17

Variables not SELECTed can equivalently be blank:

Martin Giese SPARQL, CapGemini 24 April 2020 8 / 17

Variables not SELECTed can equivalently be blank:

Martin Giese SPARQL, CapGemini 24 April 2020 8 / 17

Martin Giese SPARQL, CapGemini 24 April 2020 9 / 17

The same with blank node syntax

The same with blank node syntax

Martin Giese SPARQL, CapGemini 24 April 2020 9 / 17

The same with blank node syntax

The same with blank node syntax

▶ Groups may include constraints or filters

Martin Giese SPARQL, CapGemini 24 April 2020 10 / 17

▶ Groups may include constraints or filters

Martin Giese SPARQL, CapGemini 24 April 2020 10 / 17

Martin Giese SPARQL, CapGemini 24 April 2020 11 / 17