Querying from the command line

On this page

    The default Comunica query engine that exposes most standard features is Comunica SPARQL, which uses the package name @comunica/query-sparql. In this guide, we will install it globally, and show how it can be invoked from the command line.

    This guide discusses querying over Linked Data on the Web, exposed via URLs. Click here to query over local RDF files instead.
    Watch part of this guide in action live within this Webinar recording.

    1. Installation

    Since Comunica runs on Node.js, make sure you have Node.js installed on your machine.

    Next, we can install Comunica SPARQL on our machine:

    $ npm install -g @comunica/query-sparql

    2. SPARQL querying over one source

    After installing Comunica SPARQL, you will be given access to several commands including comunica-sparql, which allows you to execute SPARQL queries from the command line.

    This command requires one or more URLs to be provided as sources to query over. As last argument, as SPARQL query string can be provided.

    For example, the following query retrieves the first 100 triples from DBpedia:

    $ comunica-sparql https://fragments.dbpedia.org/2016-04/en \
        "SELECT * WHERE { ?s ?p ?o } LIMIT 100"
    Given a URL, Comunica will automatically detect the type of source and handle it accordingly.

    As output, a JSON array of bindings for the selected variables will be returned:


    3. Query file input

    Since SPARQL queries can sometimes become very large, it is possible to supply them via a local file using the -f option.

    Assuming a file path/myquery.sparql exists, we can query over it as follows:

    $ comunica-sparql https://fragments.dbpedia.org/2016-04/en -f path/myquery.sparql

    4. SPARQL querying over multiple sources

    One key feature of Comunica is its ability to query over multiple sources. For this, you can just supply any number of URLs as arguments. Just make sure that the last argument remains your query.

    $ comunica-sparql https://fragments.dbpedia.org/2016-04/en \
        https://www.rubensworks.net/ \
        https://ruben.verborgh.org/profile/ \
        "SELECT * WHERE { ?s ?p ?o } LIMIT 100"


    Next to SPARQL SELECT queries, it is also possible to execute CONSTRUCT queries to produce RDF triples:

    $ comunica-sparql https://fragments.dbpedia.org/2016-04/en \
        "CONSTRUCT WHERE { ?s ?p ?o } LIMIT 100"
    <http://0-access.newspaperarchive.com.topcat.switchinc.org/Viewer.aspx?img=8716084> <http://dbpedia.org/property/accessdate> "2010-04-21"^^<http://www.w3.org/2001/XMLSchema#date>;
        <http://dbpedia.org/property/date> "1939-01-02"^^<http://www.w3.org/2001/XMLSchema#date>;
        <http://dbpedia.org/property/format> "PDF";
        <http://dbpedia.org/property/isCitedBy> <http://dbpedia.org/resource/List_of_Attorneys_General_of_Wisconsin>;
        <http://dbpedia.org/property/location> "Sheboygan, Wisconsin";
        <http://dbpedia.org/property/page> "1";

    ASK queries will produce a boolean output:

    $ comunica-sparql https://fragments.dbpedia.org/2016-04/en \
        "ASK { ?s ?p ?o }"

    6. Changing result format

    SELECT queries will be printed as JSON by default, and CONSTRUCT queries as RDF TriG. This can be overridden using the -t option.

    For example, displaying results as SPARQL JSON results:

    $ comunica-sparql https://fragments.dbpedia.org/2016-04/en \
        "SELECT * WHERE { ?s ?p ?o } LIMIT 100" \
        -t 'application/sparql-results+json'
    {"head": {"vars":["s","p","o"]},
    "results": { "bindings": [
    {"s":{"value":"http://0-access.newspaperarchive.com.lib.utep.edu/us/mississippi/biloxi/biloxi-daily-herald/1899/05-06/page-6?tag=tierce+wine&rtserp=tags/tierce-wine?page=2","type":"uri"},"p":{"value":"http://dbpedia.org/property/newspaper","type":"uri"},"o":{"value":"Biloxi Daily Herald","type":"literal"}},
    {"s":{"value":"http://0-access.newspaperarchive.com.lib.utep.edu/us/mississippi/biloxi/biloxi-daily-herald/1899/05-06/page-6?tag=tierce+wine&rtserp=tags/tierce-wine?page=2","type":"uri"},"p":{"value":"http://dbpedia.org/property/title","type":"uri"},"o":{"value":"A New System of Weights and Measures","type":"literal"}},
    All available formats can be printed via comunica-sparql --listformats

    7. Printing the query plan

    Using the --explain option, the query plan can be printed via different explain modes.

    8. Learn more

    This guide only discussed the basic functionality of comunica-sparql. You can learn more options by invoking the help command:

    $ comunica-sparql evaluates SPARQL queries
    Recommended options:
      -q, --query       Evaluate the given SPARQL query string                                                                                              [string]
      -f, --file        Evaluate the SPARQL query in the given file                                                                                         [string]
      -i, --inputType   Query input format (e.g., graphql, sparql)                                                                      [string] [default: "sparql"]
      -t, --outputType  MIME type of the output (e.g., application/json)                                                                                    [string]
      -c, --context                 Use the given JSON context string or file (e.g., config.json)                                                           [string]
          --to                      Destination for update queries                                                                                          [string]
      -b, --baseIRI                 base IRI for the query (e.g., http://example.org/)                                                                      [string]
      -d, --dateTime                Sets a datetime for querying Memento-enabled archives                                                                   [string]
      -l, --logLevel                Sets the log level (e.g., debug, info, warn, ...)                                                     [string] [default: "warn"]
          --lenient                 If failing requests and parsing errors should be logged instead of causing a hard crash                                [boolean]
      -v, --version                 Prints version information                                                                                             [boolean]
          --showStackTrace          Prints the full stacktrace when errors are thrown                                                                      [boolean]
          --httpTimeout             HTTP requests timeout in milliseconds                                                                                   [number]
          --httpBodyTimeout         Makes the HTTP timeout take into account the response body stream read                                                 [boolean]
          --httpRetryCount          The number of retries to perform on failed fetch requests                                                               [number]
          --httpRetryDelay          The number of milliseconds to wait between fetch retries                                                                [number]
          --httpRetryOnServerError  If fetch should be retried on 5xx server error responses, instead of being resolved.                                   [boolean]
          --unionDefaultGraph       If the default graph should also contain the union of all named graphs                                                 [boolean]
          --noCache                 If the cache should be disabled                                                                                        [boolean]
      -p, --proxy                   Delegates all HTTP traffic through the given proxy (e.g. http://myproxy.org/?uri=)                                      [string]
          --listformats             Prints the supported MIME types                                                                                        [boolean]
          --explain                 Print the query plan                                                         [string] [choices: "parsed", "logical", "physical"]
          --localizeBlankNodes      If blank nodes should be localized per bindings entry                                                                  [boolean]
      -r, --recoverBrokenLinks      Use the WayBack machine to recover broken links                                                       [boolean] [default: false]
      comunica-sparql https://fragments.dbpedia.org/2016-04/en -q 'SELECT * { ?s ?p ?o }'
      comunica-sparql https://fragments.dbpedia.org/2016-04/en -f query.sparql
      comunica-sparql https://fragments.dbpedia.org/2016-04/en https://query.wikidata.org/sparql ...
      comunica-sparql hypermedia@https://fragments.dbpedia.org/2016-04/en sparql@https://query.wikidata.org/sparql ...