Parsed files in final too small

[kurzum]

I managed to derive mappings and wikidata.
However, after the last 5 commits or so the files are only partially generated for generic.

Setup on generic server:

# pom is modified to only derive geo-coordinates/2018.08.01
cd /data/derive/databus-maven-plugin/dbpedia/generic
mvn databus-derive:clone -e

result:

root@dbpedia-generic:/data/derive/databus-maven-plugin/dbpedia/generic/target/databus/derive# 
du -sh *
90M     downloads
3.2M    final
208K    reports

I removed the lbzip2file, but that is not it, still smaller in final:

du -sh *
90M     downloads
55M     final
508K    reports

Potential solutions:

1. I tried to revert to 4cb41bc
   4cb41bc , but the error is in there.

2. I noticed this line: https://github.com/dbpedia/databus-derive/blob/master/src/main/scala/org/dbpedia/databus/derive/mojo/CloneGoal.scala#L191
   > instead of >> but this only affects the reports, I tested it, it somehow makes bigger reports

  du -sh *
90M     downloads
3.2M    final
508K    reports

3. the files in final parse well with rapper so they seem parsed
4. A previous commit changed, but I am not sure, what this even does

val jenaTriples = new LangNTriplesSkipBad(tokenizer, parserProfile, null).filter(

-           wrappedTriple => { ! removeWarnings || ! reports.getCorruptRows.contains(wrappedTriple.getRow) }
+          wrappedTriple => { ! removeWarnings || ! errorHandler.getViolatedRowsBuffer.contains(wrappedTriple.getRow) }

5. might be the outputstream of the finalfile, but this seems straightforward

Overall, I didn't find anything useful really.