Load data into named graphs

Upload local RDF files to a workspace where you have write access. Select the workspace first, and choose the destination branch for each import.

Choose a format

Upload one file

Set a recognizable graph IRI for a triple-format file:

legra graph upload contacts.ttl --workspace research --graph https://example.org/contacts --target-branch main

Without --graph, graph names are derived from file paths. Do not assume that an upload becomes part of the default graph. For N-Quads and TriG, preserve the graph names in the input instead of supplying a single target graph.

Upload a directory

Preview the input before uploading a collection of files:

legra graph upload ./datasets/ --dry-run

Then set the graph-name prefix and destination:

legra graph upload ./datasets/ --graph-base https://example.org/datasets/ --workspace research --target-branch review

Directory uploads include supported files in subdirectories. Use Branching to create review before targeting it. Uploads replace the corresponding graph data, so review a repeat import before promoting its changes to shared data.

Verify the result

List graphs containing data:

legra graph query 'SELECT DISTINCT ?graph WHERE { GRAPH ?graph { ?s ?p ?o } }' --workspace research --branch review

For a particular graph, count its statements:

SELECT (COUNT(*) AS ?count) WHERE {
  GRAPH <https://example.org/contacts> { ?s ?p ?o }
}

Run verification against the same workspace and destination branch used for the upload. See Upload options for command details or HTTP API for graph imports over HTTP.