I want to upload an rdf graph to virtuoso from python using ld_dir of isql. I add the path in the “DirsAllowed” inside the “virtuoso.ini” config file. Then I run “ld_dir(,)” and “rdf_loader_run()” to upload the file.
I want to know how can I do the same thing but from remote location. How should I configure the “virtuoso.ini” to allow paths from a remote computer?
To use the Virtuoso RDF Bulk Loader ie “ld_dir(,)” and “rdf_loader_run()” etc, the RDF dataset files must exist be readable directly from the machine Virtuoso is running (via local ideally or virtual file system mount) for optimum performance of the bulk loading process, which is the purpose of the bulk loader.
What is the number of dataset files and total size in terms of number of triples you would be seeking to load from a remote location ?
Thank you for your replay,
I want to load a graph of 13 millions triples.
I have already tried sparqlwrapper but I found that it is not adapted for large graphs.
I have also tried the SparqlStore of rdflib library. However, it does not support blank-nodes.
Any other idea?
What is the format of these 13million triples, as I assume they are in one or more file in some standard RDF triple format ? Why can’t these RDF dataset files be loaded on the server machines Virtuoso is running on an the RDF Bulk Loader used for loading ?
Do you not have access to the server machines hosting Virtuoso ?
Do you know if the Virtuoso server has been tuned for hosting large RDF datasets based on the available machine resources ie memory for hosting them in memory for best performance ?
Note also when loading large RDF datasets on client or server side it is recommended the Virtuoso default transaction/commit mode be changed from the default of manual commit to auto-commit with the log_enable() function set to 2 or 3.