In order to perform the following steps, you need a Google Storage as well as a BigQuery account. Let us further assume that you've checked out the source from https://bigquery-linkeddata.googlecode.com/hg/ into the $bgl directory, which we will use as the base directory in the following. Step 1: Prepare the data file You'll need an RDF file in NTriples format, either from the test/ directory or grab one from the Web (for example enter 'rdf filetype:nt' into Google). Once you have the RDF file in a local directory (we assume it's in test/), run the following command (in $bgl): python tools/nt2csv.py test/mhausenblas-foaf.nt which creates the file test/mhausenblas-foaf.csv we will use in the next step. Step 2: Upload the data file You upload the data file to Google Storage into one of your buckets, shown as 'mybucket' in the following; replace it with the actual name of your bucket. gsutil cp test/mhausenblas-foaf.csv gs://mybucket/in/mhausenblas-foaf.csv Step 3: Create the RDF table To create the table that holds the RDF data, you need to specify a table schema: bq create mybucket/rdftable schema/quintuple.scheme The layout of the RDF table looks like the following: +--------------------------------------------------------+ | graph_uri | subject | predicate | object | object_type | +--------------------------------------------------------+ | ... | ... | ... | ... | ... | Step 4: Import the RDF data into the table Next you have to import the earlier created CSV file into the table: bq import mybucket/rdftable mybucket/in/mhausenblas-foaf.csv Importing can take up to 10 minutes if you have a lot of data. Step 5: Run a query Once the CSV file is imported into the table, we can run a query: $ bq query "SELECT object FROM [mybucket/rdftable] WHERE predicate = 'http://xmlns.com/foaf/0.1/knows' LIMIT 10" The query above lists ten people that I know, which corresponds to the following SPARQL query: PREFIX foaf: SELECT ?o FROM WHERE { ?s foaf:knows ?o . } LIMIT 10