Summer in Greece, vol.2

问题

Following this qeustion: Summer in Greece with SPARQL, since my memory runs out when executing this query, I would like to constrain the query between two regional units, but I can't group them:

SELECT * #?municipality (?bwCount1+?bwCount2 as ?bwCount)
WHERE { 
{
    SELECT (COUNT(?bw) as ?bwCount1) WHERE 
    {
           ?regional_unit geo:έχει_επίσημο_όνομα "ΠΕΡΙΦΕΡΕΙΑΚΗ ΕΝΟΤΗΤΑ ΗΡΑΚΛΕΙΟΥ" .
           ?municipality1 geo:ανήκει_σε ?regional_unit .
           ?municipality1 geo:έχει_γεωμετρία ?geometry .

           ?bw geos:hasGeometry ?bw_geo .
           ?bw_geo geos:asWKT ?bw_geo_wkt .
           FILTER(strdf:within(?geometry, ?bw_geo_wkt)) .
           ?bw unt:has_concie_0 ?concie_0 .
           FILTER(?concie_0 > 40)
    }
}
UNION
{
    SELECT (COUNT(?bw) as ?bwCount2) WHERE 
    {
           ?regional_unit geo:έχει_επίσημο_όνομα "ΠΕΡΙΦΕΡΕΙΑΚΗ ΕΝΟΤΗΤΑ ΛΕΣΒΟΥ" .
           ?municipality2 geo:ανήκει_σε ?regional_unit .
           ?municipality2 geo:έχει_γεωμετρία ?geometry .

           ?bw geos:hasGeometry ?bw_geo .
           ?bw_geo geos:asWKT ?bw_geo_wkt .
           FILTER(strdf:within(?geometry, ?bw_geo_wkt)) .
           ?bw unt:has_concie_0 ?concie_0 .
           FILTER(?concie_0 > 40)
    }
}
}
#GROUP BY ?municipality
#ORDER BY DESC(?bwCount)

What am I missing?

回答1:

You may again be confusing yourself with the sub-selects, and this could be causing inefficient use of memory. I see a couple of way of getting the results I believe you want (it is a bit unclear). The first is by using UNION:

SELECT (COUNT(?bw1) as ?bwCount1) (COUNT(?bw2) as ?bwCount2)
WHERE { 
   ?municipality1 geo:ανήκει_σε ?regional_unit .
   ?municipality1 geo:έχει_γεωμετρία ?geometry .
   {
     ?regional_unit geo:έχει_επίσημο_όνομα "ΠΕΡΙΦΕΡΕΙΑΚΗ ΕΝΟΤΗΤΑ ΗΡΑΚΛΕΙΟΥ" .
     ?bw2 geos:hasGeometry ?bw_geo .
   }
   UNION
   {
     ?regional_unit geo:έχει_επίσημο_όνομα "ΠΕΡΙΦΕΡΕΙΑΚΗ ΕΝΟΤΗΤΑ ΛΕΣΒΟΥ" .
     ?bw1 geos:hasGeometry ?bw_geo .
   }
   ?bw_geo geos:asWKT ?bw_geo_wkt .
   FILTER(strdf:within(?geometry, ?bw_geo_wkt)) .
   ?bw unt:has_concie_0 ?concie_0 .
   FILTER(?concie_0 > 40)
}

A best practice here is to not repeat triple patterns inside the UNION unless they are part of the disjunction computation.

You could also do this with a group by on ?regional_unit:

SELECT (COUNT(?bw) as ?bwCount1)
WHERE { 
  VALUES ?name {"ΠΕΡΙΦΕΡΕΙΑΚΗ ΕΝΟΤΗΤΑ ΗΡΑΚΛΕΙΟΥ" "ΠΕΡΙΦΕΡΕΙΑΚΗ ΕΝΟΤΗΤΑ ΛΕΣΒΟΥ"}
  ?municipality1 geo:ανήκει_σε ?regional_unit .
  ?municipality1 geo:έχει_γεωμετρία ?geometry .
  ?regional_unit geo:έχει_επίσημο_όνομα ?name .
  ?bw geos:hasGeometry ?bw_geo .
  ?bw_geo geos:asWKT ?bw_geo_wkt .
  FILTER(strdf:within(?geometry, ?bw_geo_wkt)) .
  ?bw unt:has_concie_0 ?concie_0 .
  FILTER(?concie_0 > 40)
} GROUP BY ?regional_unit

来源：https://stackoverflow.com/questions/36530519/summer-in-greece-vol-2

标签

group-by

sparql

rdf

geospatial

linked-data