Hi all,
I'm trying to load somehow Billion NQUADS to Stardog (the latest version running on Ubuntu).
I've set the Stardog server with strict.parsing=false. I want to skip errors during loading time.
However, I just got there error:
"There was a fatal failure during preparation of 1706d549-142f-49eb-9b4f-717286541a60 com.stardog.stark.io.InvalidRDF: No separator character found in the URI: iri_id_371642758_with_no_name_entry "
Is it possible to solve skip this issue and continue the loading process?
TIA
Best,
Ghislain
Just to close the loop on this for others reading the forum, the strict.parsing should be set as a database option and setting it inside stardog.properties does not have an impact.
I was under the impression that strict.parsing only applied to literals. Does it apply to IRI's as well?
You could use the rapper parser from the redland library to loop over each line of your NQUADS with a bash script and parse it as though it was a NQUAD with a single triple, then dump the bad ones to the file and optionally include the line number. Raptor would allow you to avoid the JVM start up for each QUAD. I don't know how long it would take for a billion but you could easily parallelize it by breaking it into chunks of 1m.