2018-01-20 11:53 GMT+01:00 Andy Seaborne <[email protected]>: > > On 20/01/18 08:37, Jean-Marc Vanel wrote: > >> 2018-01-20 0:15 GMT+01:00 Andy Seaborne <[email protected]>: >> >> Thanks for the clear explanations. >> This is indeed flexible but a bit time consuming . >> Ideally, a new option in riot command line would say: recover and continue >> after errors. Want an issue ? >> > > Even if you get bad URIs in, there will be problems because they can't > necessarily write them out legally. > > Much, much better to correct/fixup the data before it goes in. Clean-up > has to happen at some point. > > There are many ways URI can go wrong, it isn't just a case of saying > "accept any string", for example, ">" in a URI can't be written in > N-Triples even in some lax output mode. > > Andy >
I've been warned ! Even if I'm at the end of data flow, I did my best to clean stuff, reusing Conal's code: https://framagit.org/Scrutari/RDFexport/blob/master/scrutari-to-rdf.xslt#L133 Note that this XSLT: $string castable as xs:anyURI catches almost nothing . In TDB, I will replace the graph containing the former content with the RDF generated with the latest XSLT . I never had troubles with DROP GRAPH . -- Jean-Marc Vanel http://www.semantic-forms.cc:9111/display?displayuri=http://jmvanel.free.fr/jmv.rdf%23me#subject <http://www.semantic-forms.cc:9111/display?displayuri=http://jmvanel.free.fr/jmv.rdf%23me> Déductions SARL - Consulting, services, training, Rule-based programming, Semantic Web +33 (0)6 89 16 29 52 Twitter: @jmvanel , @jmvanel_fr ; chat: irc://irc.freenode.net#eulergui
