DESCRIBE of two resources causes Cartesian Product

[VladimirAlexiev]

Current Behavior

We're trying to get a description of an ontology and all terms within with this query:

PREFIX swapi: <https://swapi.co/ontology/>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
DESCRIBE swapi: ?term { ?term rdfs:isDefinedBy swapi: }

The result is attached: query-result.txt (rename to .nt):

• wc -l query-result.nt
  15087 total lines
• sort query-result.nt|uniq|wc -l
  2280 unique lines (7x duplication on average)
• sort query-result.nt|uniq -c|grep -v " 1 "|sort -r -n|less
  to view the top repeated lines
• sort query-result.nt|uniq -c|grep " 114 "|grep -v isDefinedBy
  to see that the most repeated lines (114x) are all rdfs:isDefinedBy
• sort query-result.nt|uniq -c|grep " 114 "|sort|uniq|wc -l
  113: to see that there are 113 terms, and so the duplication is caused by some Cartesian Product that produces o(N^2) results
• sort query-result.nt|uniq -c|grep " 2 "|less
  to see that the other duplicated lines are of the form
  <https://swapi.co/vocabulary/Clawdite> a <https://swapi.co/vocabulary/Species>

Note: The ontology is a bit unusual since it has a "metaclass" Character and then per-species classes that describe typical characteristics of that species, see below.
But I think that's beside the point since the major duplication comes from DESCRIBE swapi: ?term

# grep -i clawdite query-result.nt|riot --formatted ttl
<https://swapi.co/resource/clawdite/70>
        a       <https://swapi.co/vocabulary/Clawdite> , <https://swapi.co/vocabulary/Character> .

<https://swapi.co/vocabulary/Clawdite>
        a       <https://swapi.co/vocabulary/Species> , <http://www.w3.org/2002/07/owl#Class> ;
        <http://www.w3.org/2000/01/rdf-schema#isDefinedBy>
                <https://swapi.co/ontology/> ;
        <http://www.w3.org/2000/01/rdf-schema#label>
                "Clawdite" ;
        <http://www.w3.org/2000/01/rdf-schema#subClassOf>
                <https://swapi.co/vocabulary/Sentient> , <https://swapi.co/vocabulary/Reptilian> ;
        <https://swapi.co/vocabulary/averageHeight>
                180.0 ;
        <https://swapi.co/vocabulary/averageLifespan>
                "70" ;
        <https://swapi.co/vocabulary/character>
                <https://swapi.co/resource/clawdite/70> ;
        <https://swapi.co/vocabulary/eyeColor>
                "yellow" ;
        <https://swapi.co/vocabulary/film>
                <https://swapi.co/resource/film/5> , <https://swapi.co/resource/film/6> ;
        <https://swapi.co/vocabulary/language>
                "Clawdite" ;
        <https://swapi.co/vocabulary/planet>
                <https://swapi.co/resource/planet/54> ;
        <https://swapi.co/vocabulary/skinColor>
                "yellow" , "green" .

Expected Behavior

No duplicates, or some triples duplicated only twice (eg each isDefinedBy from the point of view of subject and object)

Steps To Reproduce

• load SWAPI turtle: swapi.txt
• Run the above query
• Examine the results, or save to ntriples and repeat the analysis above using shell commands

Version

rdf4j 4.3.8 (GraphDB 10.4.2)

Are you interested in contributing a solution yourself?

None

Anything else?

No response