search

RMLio / rmlmapper-java

The RMLMapper executes RML rules to generate high quality Linked Data from multiple originally (semi-)structured data sources
http://rml.io
MIT License
144 stars 61 forks source link

RMLMapper v6.5.1 doesn't generate all triples #234

Closed dachafra closed 1 month ago

dachafra commented 3 months ago

We tested a very simple mapping and data with the combination of two columns for the subject generation and the last version of the engine generates 852 triples, and when we run it with the v6.2.2 generates 563082 triples (93417 if we remove the duplicates).

We attached the mapping and the data we are using.

/cc @OscarHid

resources.zip

goanes commented 1 month ago

Hi I have the same issue with this warning : I use the docker:latest version (as of decembre 2023)

09:18:49.202 [main] WARN  b.u.idlab.knows.dataio.record.Record.<init>(20) - Header has more columns than this row
09:18:49.340 [main] WARN  b.ugent.rml.functions.ConcatFunction.concat(68) - Not all values for a template where found. More specific, the variable ReferenceExecutor that works with hosts did not provide any results.

I have tried to roll back to the version 6.2.2 that you use and I have no warning.

DylanVanAssche commented 1 month ago

7.0.0 was released which refactors a lot, could you please try it? Thanks!

DylanVanAssche commented 1 month ago

I tested this with development and got 93417 triples, closing as this is what is expected :)