12 million triples is not at all large for Fuseki/TDB. It may or may not be very large for inference depending on your entailment regime and ontology, but that wouldn't seem to give the error you showed.
>> The data in DBMS(mysql) cannot be wrong, right? It would seem to be as possible as data being wrong anywhere else. You can check whether there are syntax errors in your RDF data with Jena's `riot` tool. Please provide some sample data that you are actually using. If it's too large to be excerpted to this list, you can provide it as a Github gist or the like. The configuration you included printed unindented and with a lot of "@PreFix[https://github.com/prefix]"; included for some reason, which makes it very hard to read. Please either include a fair copy of your configuration or again, gist or pastebin it somewhere. ajs6f > On Dec 9, 2017, at 10:05 AM, Chan Simmer <[email protected]> wrote: > > Yes, it works well on a demo dataset. Maybe it's really the problem of my > dataset which is quite large with 12 million triples. > The data in DBMS(mysql) cannot be wrong, right? I'll check the mapping > file. > Thanks for your patience! > > > 2017-12-09 22:12 GMT+08:00 Laura Morales <[email protected]>: > >>> If there is something wrong with >>> the exported data, how can tdbloader convert it correctly >> >> It can't. It can only validate the data. >> >>> And fuseki works >>> well with the tdb without inference. >> >> But you said Fuseki also works well with inference on a different dataset? >>
