For each dataset, where should I get the actual RDF?

Contributor: 
Contributor: 
Contributor: 
Status: 
answered
Description: 
<br>
While the URIs of an Abstract Dataset take the form of, e.g., http://logd.tw.rpi.edu/source/census-gov/dataset/us_regdiv, the URIs of its versions take the form of, e.g. http://logd.tw.rpi.edu/source/census-gov/dataset/us_regdiv/version/2000-Sep-07. An Abstract Dataset does not have any associated data files because it groups the potentially many versions one may obtain from the source organization. The "actual RDF" of a Dataset can be found by selecting the value of the void:dataDump property, whose domain is a void:Dataset and whose range is a rdfs:Resource. Request that rdfs:Resource and you'll get an RDF file. For size and practicality reasons, the syntax is currently Turtle.
If you're a human, one potential path you can use to find the RDF dump file is by browsing this path:
  • Click "Datasets" on the left menu bar
    • This leads to http://logd.tw.rpi.edu/datasets
  • Click "census-gov us_regdiv" within the listing of datasets
    • This dereferences the Abstract Dataset URI http://logd.tw.rpi.edu/source/census-gov/dataset/us_regdiv
  • Under "Subsets", click http://logd.tw.rpi.edu/source/census-gov/dataset/us_regdiv/version/2000-Sep-07
  • Under "Subsets", click http://logd.tw.rpi.edu/source/census-gov/dataset/us_regdiv/version/2000-Sep-07/conversion/enhancement/1
  • Under "Dump File", select "census-gov-us_regdiv-2000-Sep-07.e1"
    • This leads to http://logd.tw.rpi.edu/source/census-gov/file/us_regdiv/version/2000-Sep-07/conversion/census-gov-us_regdiv-2000-Sep-07.e1
If you're a human writing code to get the dump files, execute the followig query at http://logd.tw.rpi.edu/sparql
prefix void: <http://rdfs.org/ns/void#>
select distinct ?dataset ?dumpFile
where {
  graph <http://logd.tw.rpi.edu/vocab/Dataset>  {
    ?dataset a void:Dataset;
             void:dataDump ?dumpFile .
  }
} 
order by ?dataset
Of course, you can get "Actual RDF" from the SPARQL endpoint if the dataset is loaded by using a CONSTRUCT query. Each dataset is in a named graph named after the dataset itself. Generally, the Versioned Datasets are loaded instead of the Abstract Datasets. The following selects the rdf:type triples from a Versioned Dataset.
prefix void: <http://rdfs.org/ns/void#>
construct {
 ?a rdf:type ?b
}
where {
  graph <http://logd.tw.rpi.edu/source/census-gov/dataset/us_regdiv/version/2000-Sep-07>  {
   ?a a ?b
  }
}