If a Wikidata entity dump is being aborted, a subsequent run should be able to pick up where it left. This will work by checking what temp files are already present, and setting the respective batch parameters based on that.

This should (probably) be usable by adding a --continue parameter to dumpwikidatajson.sh or dumpwikidatardf.sh respectively. If that is not present, the old temp files should be deleted.