I want to directly import the data of neo4j into dgraph. Is there a feasible scheme？
TL;DR: We created a new tool called Bulkloader to populate Dgraph with an initial dataset. It’s highly performant and is able to load the entire Stack Overflow dataset in just over 1 hour on a single 64 core machine. The Stack Overflow dataset is massive and consists of ~2 billion RDFs. An additional ~1 billion edges are created for indexing and other internal usages. This gives* ~3 billion edges in total. This equates to a loading rate of ~820k edges per second! *Note that we had a significant skew in the predicate distribution which slowed the load down (all the text for questions, answers and comments was on the same predicate). Better results can likely be achieved on datasets with a more even predicate distribution.