Update

QLever supports SPARQL 1.1 Update and the SPARQL 1.1 Graph Store HTTP Protocol for updating the data at runtime after building an index.

Configuration

Required: The execution of update operations requires an access token. You can set the access token via ACCESS_TOKEN in the [server] section of the QLeverfile or via the command-line option --access-token. When the server is started with --no-access-check, no access token is required.

Optional: To persist the updates across server restarts, set PERSIST_UPDATES = true in the [server] section of the QLeverfile or use the command-line option --persist-updates.

SPARQL 1.1 Update

Via HTTP

The access token can be provided either via the query parameter access-token=... or via the HTTP header Authorization: Bearer .... Here is an example using curl:

Access token as query parameterAccess token as header

curl -X "POST"
     -H "Content-Type: application/sparql-update"
     -d "DELETE WHERE { ?s ?p ?o }"
     "http://localhost:7019?access-token={token}"

curl -X "POST"
     -H "Content-Type: application/sparql-update"
     -H "Authorization: Bearer {token}"
     -d "DELETE WHERE { ?s ?p ?o }"
     "http://localhost:7019"

Via the QLever UI

You can insert the access token into the Access Token field under the Backend Information. You can use the editor to write and execute updates, analogous to how you would do for queries. In particular, syntax highlighting, formatting and basic autocompletion are available. When an access token has been entered, additional buttons for the privileged operations Reset Updates and Clear Cache (Complete) become available below the editor.

SPARQL 1.1 Graph Store HTTP Protocol

Here are some example curl commands for the different HTTP methods, extended by the non-standard TSOP method.

GETPUTDELETEPOSTTSOPHEAD

Get all triples from a graph

curl -X "GET" \
     -H "Accept: text/turtle" \
     "http://localhost:7019?graph=http://example.com/person/1.ttl"

Replace all triples in a graph

curl -X "PUT" \
     -H "Content-Type: text/turtle" \
     -H "Authorization: Bearer {token}" \
     --data-binary @graph.ttl \
     "http://localhost:7019?graph=http://example.com/person/1.ttl"

Delete all triples from a graph

curl -X "DELETE" \
     -H "Authorization: Bearer {token}" \
     "http://localhost:7019?graph=http://example.com/person/1.ttl"

Add triples to a graph

curl -X "POST" \
     -H "Content-Type: text/turtle" \
     -H "Authorization: Bearer {token}" \
     --data-binary @graph.ttl \
     "http://localhost:7019?graph=http://example.com/person/1.ttl"

Delete triples from a graph

curl -X "TSOP" \
     -H "Content-Type: text/turtle" \
     -H "Authorization: Bearer {token}" \
     --data-binary @graph.ttl \
     "http://localhost:7019?graph=http://example.com/person/1.ttl"

Same as GET, but without response body

curl -I "http://localhost:7019?graph=http://example.com/person/1.ttl"

Updates involving large files

When sending large files with curl, you should use --data-binary @file instead of --data @file or --data-urlencode @file. The first sends the data as is, the last both process the input before sending it, which leads to problems for large files.