I have a dataset that contains a large number of resources I need to reconcile against various existing data. The most straight forward approach is to do some simple string comparisons between various literals.
Unfortunately the literals are typed xsd:string, and the EulerSharp builtin str:contains isn't working on typed literals.
I have read everything I can find on the eulersharp builtins and the closest I could come is to try to cast the xsd:string to a PlainLiteral using the rdf:PlainLiteral predicate from RIF; however, this does not appear to be supported by EulerSharp.
How can I manipulate and compare Literals of type xsd:string?
Or, should I resort to preprocessing the data to strip off the datatypes?
The following .n3 file demonstrates the problem:
@prefix : <http://local#> .
@prefix str: <http://www.w3.org/2000/10/swap/string#> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
:Bob :name "Bob Smith" .
:Rob :name "Rob Smith"^^xsd:string .
{ ?P :name ?N .
?N str:contains "Smith" } => { ?P :bingo ?N } .
And the result of running the above file:
$ eye --swipl test.n3 --pass --nope
Id: euler.yap 5974 2013-02-12 00:29:00Z josd
SWI-Prolog 5.10.4 (amd64): Dec 27 2011, 08:54:16
starting 80 [msec cputime] 78 [msec walltime]
GET file:///home/ubuntu/src/test/test.n3 SC=4
networking 0 [msec cputime] 2 [msec walltime]
#Processed by Id: euler.yap 5974 2013-02-12 00:29:00Z josd
#eye --swipl test.n3 --pass --nope
@prefix : <http://local#>.
@prefix str: <http://www.w3.org/2000/10/swap/string#>.
@prefix xsd: <http://www.w3.org/2001/XMLSchema#>.
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>.
@prefix var: <http://localhost/var#>.
@prefix e: <http://eulersharp.sourceforge.net/2003/03swap/log-rules#>.
@prefix r: <http://www.w3.org/2000/10/swap/reason#>.
@prefix n3: <http://www.w3.org/2004/06/rei#>.
:Bob :name "Bob Smith".
:Rob :name "Rob Smith"^^xsd:string.
:Bob :bingo "Bob Smith".
TC=4 TP=8 BC=0 BP=5 PM=0 CM=0 FM=0 AM=0
reasoning 0 [msec cputime] 4 [msec walltime]
#ENDS 0 [msec] TC=4 TP=8 BC=0 BP=5 PM=0 CM=0 FM=0 AM=0
I asked on the EulerSharp list and received this reply:
You can get the literal from a datatyped literal via log:dtlit.
The extra rule would make it work.
Not that instead of ?dt you could also use xsd:string
@prefix : <http://local#> .
@prefix str: <http://www.w3.org/2000/10/swap/string#> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix log: <http://www.w3.org/2000/10/swap/log#>.
:Bob :name "Bob Smith" .
:Rob :name "Rob Smith"^^xsd:string .
{ ?P :name ?N .
?N str:contains "Smith" } => { ?P :bingo ?N } .
{ ?P :name ?N .
(?lit ?dt) log:dtlit ?N.
?lit str:contains "Smith" } => { ?P :bingo ?N } .
Running this gives the desired result:
$ eye --swipl test.n3 --pass --nope
Id: euler.yap 5974 2013-02-12 00:29:00Z josd
SWI-Prolog 5.10.4 (amd64): Dec 27 2011, 08:54:16
starting 50 [msec cputime] 62 [msec walltime]
#Processed by Id: euler.yap 5974 2013-02-12 00:29:00Z josd
#eye --swipl test.n3 --pass --nope
GET file:///home/ubuntu/src/test/test.n3 SC=4
networking 10 [msec cputime] 2 [msec walltime]
@prefix : <http://local#>.
@prefix str: <http://www.w3.org/2000/10/swap/string#>.
@prefix xsd: <http://www.w3.org/2001/XMLSchema#>.
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>.
@prefix log: <http://www.w3.org/2000/10/swap/log#>.
@prefix var: <http://localhost/var#>.
@prefix e: <http://eulersharp.sourceforge.net/2003/03swap/log-rules#>.
@prefix r: <http://www.w3.org/2000/10/swap/reason#>.
@prefix n3: <http://www.w3.org/2004/06/rei#>.
:Bob :name "Bob Smith".
:Rob :name "Rob Smith"^^xsd:string.
:Bob :bingo "Bob Smith".
:Rob :bingo "Rob Smith"^^xsd:string.
#ENDS 0 [msec] TC=6 TP=12 BC=0 BP=7 PM=0 CM=0 FM=0 AM=0
TC=6 TP=12 BC=0 BP=7 PM=0 CM=0 FM=0 AM=0
reasoning 0 [msec cputime] 2 [msec walltime]