returns a lot of types for Victor Hugo (cf url below), whereas a grep in “instance-types_lang=en_transitive.ttl.bz2” only returns 5 (cf screenshot). Which file from DBpedia contains the rest of the types of Victor Hugo?
I grep for http://dbpedia.org/resource/Victor_Hugo in these 3 datasets and in total I get 99 statements, which is equal to what you get via the SPARQL endpoint.
In summary, the SPARQL endpoint loads a specific collection of datasets (not all) that are extracted by DBpedia extraction framework or hosted at the DBpedia Databus platform.
I have had a look inside the Yago and SDTypes directories.
The files have pretty recent timestamps.
Why not change the directory names to reflect that freshness?
(2016.10.01/ is misleading)
A nice-to-have proposal:
Currently I need to update a live instance of DBPedia from 9 months ago, and I cannot retrieve the corresponding Yago and SDTypes files.
It is probably a good idea to put the different iterations of the Yago and SDTimes files in separate timestamped directories. [Just as you do for the instance-types file, and most of the other files of DBPedia].