Github user afs commented on the issue:
https://github.com/apache/jena/pull/426
Long term (!), I'd like to get some aspects TDB1 `tdbloader2` into the TDB2
but in Java for portability. `tdbloader2` is breaking up the task of building a
single index into machine-resource sized chunks. `LoaderMain` does not consider
multiple tasks to build one index; this also means `LoaderPlan`s are not
completely stable as an API.
In doing work on loading, I can see more possibilities which need
experimentation to see if they are a good idea or a mad idea. This PR is
breaking the "wait until perfect" effect. This loader should be good for a
100's millions triples (I haven't tested a billion yet) so it is a significant
step forward in itself.
---
