Filter: Evaluable Expression. List Syntax. Tests. Filter: Evaluable Expression. Tests : functions. Tests 06/09/2013. (1?x v )

Transcription

1 SPARQL 2 W3C RDF Data Access SPARQL W3C Simple Protocol And RDF Query Language olivier.corby@inria.fr Use Case and Requirements : Query language : XML Result Format: Protocol for Web Services : SPARQL 3 SPARQL : example 4 Semantic Web SQL PREFIX foaf: < Query an RDF triple store as graph(s) SPARQL is neutral w.r.t inferences RDF/RDFS/OWL/RIF (e.g. subsumption, transitivity, etc.) SELECT?mbox WHERE {?x foaf:name Olivier.?x foaf:mbox?mbox The triple store may process inferences Return the mail box of resources whose name is Olivier SPARQL 1.0 does not enable to tune inferences SPARQL Syntax: Triple 5 Blank Node 6 subject property object < foaf:name?name?x c:speed [ rdf:value?val ; c:unit 'km/h' ]?x foaf:name 'Olivier' Equivalent to :?x?p 'Olivier' _:b1 foaf:age 49?x c:speed _:b _:b rdf:value?val _:b c:unit 'km/h' 1

2 List Syntax 7 Filter: Evaluable Expression 8 (1?x v ) _:a rdf:first 1. _:a rdf:rest _:b. _:b rdf:first?x. _:b rdf:rest _:c. _:c rdf:first v. _:c rdf:rest rdf:nil FILTER (?age >= 7 &&?age <= 77) Comparison : < <= = >= >!= Operation : + * / - Boolean : && (and) (or)! (not) Function : isblank(?x) my:fun(?y) Filter: Evaluable Expression 9 Tests 10?x c:age?age Match a regular expression FILTER (?age >= 7 &&?age <= 77) Variable must be bound by a triple regex(?string,.*inria.* ) Tests 11 Tests : functions 12 Compare literals with and without language isuri(?x) isliteral(?y) engineer = str( isblank(?z) bound(?t) 2

3 Tests : cast 13 Optional Pattern 14 xsd:integer(?x) >= 18 xsd:date( ) xsd:string( ) xsd:double() xsd:boolean() SELECT * WHERE {?x :hascreated?doc. OPTIONAL {?x :ismemberof?org hascreated is mandatory ismemberof is optional Optional Pattern 15 Optional Pattern 16 SELECT * WHERE {?x :hascreated?doc. OPTIONAL {?x :age?age.?x :ismemberof?org SELECT * WHERE {?x :hascreated?doc. OPTIONAL {?x :age?age OPTIONAL {?x :ismemberof?org age and ismemberof must be present for the success of optional Pattern UNION 17 RDF Dataset 18 SELECT * WHERE { {?x :apply?job UNION {?x :practice?act :Francis :apply :Presidency :Nick :practice :Bike :Nick :apply :Presidency Results (1) x = :Francis ; job = :Presidency ; act = unbound (2) x = :Nick ; job = unbound ; act = :Bike (3) x = :Nick ; job = :Presidency ; act = unbound Query an RDF base with several graphs Named graphs with URIs Default graph Identify the graphs that are queried 3

4 SPARQL Graph Pattern SPARQL Graph Pattern g1 Olivier INRIA Olivier INRIA Corese - 19 Corese - 20 SPARQL Graph Pattern SPARQL Graph Pattern select?g graph?g {?x?org g1 g2 g1 g2 Olivier INRIA James CNRS Olivier INRIA James CNRS Corese - 21 Corese - 22 SPARQL Graph Pattern select?g graph?g {?x?org SPARQL Graph Pattern g1 Olivier INRIA James CNRS g2 g1 Olivier (1)?g = g1 (2)?g = g2 INRIA James CNRS Corese - 23 g2 select * from named <g1> from named <g2> graph?g {?x?org Corese

5 g1 Olivier SPARQL Default Graph INRIA select * from <g1> from <g2>?x?org James CNRS Corese - 25 g2 Result SELECT * WHERE SELECT DISTINCT?x?y WHERE ORDER BY?x DESC(?y) LIMIT 10 OFFSET Distinct 27 Order by 28 select distinct?x?z?x :friend?y?y :friend?z Eliminate duplicate results (eliminate same variables with same values) select?pers?date :Jim :author :d2 :Jack :author :d1?pers :author?doc :d2 :date ?doc :date?date :d1 :date order by?date Result (1) pers = :Jim ; date = (2) pers = :Jack ; date = Limit 29 Limit/Offset 30 select * PATTERN LIMIT 20 Return 20 results at most select * PATTERN LIMIT 20 OFFSET results (at most), after the 10th result i.e. from 11th to 30th 5

6 Construct Return new RDF graph construct {?girl :hasbrother?boy?boy :hassister?girl 31 Describe describe?x?x rdf:type :HotTopic?x :date?date filter (?date >= ^^xsd:date) Return a description of?x The appropriate description of a resource is determined by the RDF store 32 Ask 33 Negation as failure 34 Persons that are member of an organization and who have not created a document : ask {:Olivier :teach :ensi yes/no answer SELECT * WHERE {?x c:ismemberof?org. OPTIONAL {?x c:hascreated?doc FILTER (! bound(?doc)) Extension Function (user defined) PREFIX fun: < SELECT * WHERE {?x c:age?age. FILTER fun:prime(?age) 35 XML Result Format <?xml version="1.0"?> <sparql xmlns=" <head> <variable name="x"/> <variable name="hpage"/> <variable name="name"/> <variable name="age"/> <variable name="blurb"/> </head>... </sparql> 36 6

7 XML Result Format <?xml version="1.0"?> <sparql xmlns=" <head>... </head> <results> <result>... </result> <result>... </result> </results> </sparql> 37 XML Result Format <result> <binding name="x"><bnode>r2</bnode></binding> <binding name="hpage"><uri> <binding name="name"> <literal xml:lang="en">bob</literal> </binding> <binding name="age"> <literal datatype= &rdfs;integer">30</literal> </binding> </result> 38 SPARQL 1.1 Query Language 39 Project Expression 40 W3C Project Expression New statements Minus, Exists Subquery Return the result of an expression select * (ext:price(?doc) as?price)?doc rdf:type c:document?doc c:author?a New filters 41 New functions 42 coalesce(?x,?y, 10): return first value that is not an error (such as unbound) strlen, concat, substr, strstarts, strends, contains if(?x>10,?y,?x+10)?x in ( alpha,?beta, gamma ) year, month, day, hours, minutes, seconds 7

8 43 44 Compute a global result over the list of results: min, max, count sum, avg group_concat sample select(count(?doc) as?count)?x c:hascreated?doc 45 : distinct 46 select?x (count(?doc) as?count)?x c:hascreated?doc group by?x select?x (count(distinct?doc) as?count)?x foaf:knows?y?y c:hascreated?doc group by?x Having Additional filter after aggregate select?x (count(?doc) as?count)?x c:hascreated?doc group by?x having (count(?doc) >= 10) 1. min 2. max 3. sum 4. avg 5. count 6. group_concat 7. sample 8

9 49 50 Return one result when there is no group by Return one result for each «group by» key select (min(?price) as?min)?x ex:price?price select?class (min(?price) as?min)?x a?class ; ex:price?price group by?class Count the number of results Count the number of distinct results select (count(*) as?count)?x ex:price?price select (count(distinct *) as?count)?x ex:price?price Retrieve resources linked by a path in the graph Retrieve resources linked by a path in the graph?x foaf:knows?y?x foaf:knows/foaf:knows?y 9

10 55 56 Retrieve resources linked by a path in the graph?x foaf:knows/foaf:knows/foaf:knows?y prefix foaf: < > select *?x foaf:knows+?y ; foaf:name Olivier prefix foaf: < > select *?x foaf:knows+?y ; foaf:name Olivier foaf:knows Olivier Fabien prefix foaf: < > select *?x foaf:knows+?y ; foaf:name Olivier foaf:knows Olivier Fabien foaf:knows Alain 59 first _:b1 am 60 prefix foaf: < > select *?x foaf:knows+?y ; foaf:name Olivier foaf:knows Olivier Fabien foaf:knows foaf:knows Alain Oumy rest _:b2 first stram rest _:b3 first gram select?val xxx rdf:rest*/rdf:first?val 10

11 Expression Operators 61 Negation : Minus 62 / : sequence : alternative + : one or several Remove from the member of org the resources whose name is Olivier * : zero or several? : optional ^ : reverse! : negation select *?x c:memberof?org minus {?x c:name Olivier MINUS may remove nothing 63 Negation : Not Exists 64 select *?x c:memberof?org minus {?y c:name Olivier?x c:memberof?org. filter( not exists {?x c:author?doc ) Remove results that are compatible (same variables have same values) when there is at least one common variable Minus vs Exists 65 Minus vs Exists 66 Same results: Different results:?x c:memberof?org. filter(! exists {?x c:author?doc )?x c:memberof?org. filter(! exists {?y c:author?doc )?x c:memberof?org. minus {?x c:author?doc?x c:memberof?org. minus {?y c:author?doc 11

12 Sub Query 67 Bindings 68 Find properties of the cheapest car select * {select (min(?price) as?min)?car ex:hasprice?price?car ex:hasprice?min?car?p?val select *?x foaf:knows?y values?x { ex:robert ex:jimmy ex:john Service 69 SPARQL 1.1 Update 70 select *?x foaf:name Olivier service < {?x?p?y Update language for RDF graphs CRUD: Create Read Update Delete Insert Data 71 Delete Data 72 PREFIX dc: < PREFIX dc: < INSERT DATA { DELETE DATA { < dc:title "A new book" ; < dc:title "A new book" ; dc:creator "A.N.Other". dc:creator "A.N.Other". 12

13 Delete 73 Insert 74 PREFIX foaf: < PREFIX foaf: < DELETE {?person foaf:firstname 'Bill' INSERT {?person foaf:firstname 'William' WHERE {?person a foaf:person.?person foaf:firstname 'Bill' WHERE {?person a foaf:person.?person foaf:firstname 'Bill' Delete Insert 75 Insert Graph 76 PREFIX foaf: < PREFIX foaf: < DELETE {?person foaf:firstname 'Bill' INSERT {?person foaf:firstname 'William' WHERE {?person a foaf:person.?person foaf:firstname 'Bill' INSERT { graph <g1> {?person foaf:firstname 'William' WHERE {?person a foaf:person.?person foaf:firstname 'Bill' Using: like From 77 With 78 PREFIX foaf: < PREFIX foaf: < INSERT {?person foaf:firstname 'William' WITH <g1> USING <g1> INSERT {?person foaf:firstname 'William' WHERE {?person a foaf:person.?person foaf:firstname 'Bill' WHERE {?person a foaf:person.?person foaf:firstname 'Bill' 13

14 Update LOAD <documenturi> [ INTO GRAPH <uri> ] 79 CLEAR [ SILENT ] (GRAPH <uri> DEFAULT NAMED ALL ) DROP [ SILENT ] (GRAPH <uri> DEFAULT NAMED ALL ) CREATE [ SILENT ] GRAPH <uri> 14