On Saturday, March 12, 2022 at 11:28:32 AM UTC Nicholas Car wrote: > Any feedback on this format or on RDF text files and version control in > general would be great. >
Although It’s not advertised as such, the architecture of the BerkeleyDB RDFLib Store implementation preserves the order in which triples are added (as a a side-effect of key indexing) and so all of the RDFLib Stores based on key-value back-ends (BerkelyDB, LevelDB, SQLiteLSM) provide reliably repeatable re-serialization suitable for unambitious efforts. I've not yet checked but I imagine that the corresponding (again, index-using) AbstractSQL-based RDFLib Stores (atm, only rdflib-sqlalchemy) also exhibit the same repeatable serialization property. Otherwise, if you're just interested in manageable diffs of large but straightforward graphs, serializing as ntriples/nquads and sorting the serialization is also a viable strategy, I've found <https://github.com/DOACC/individuals/commit/1515db20e936a085d36a40210cb63237b6f2e837>. However this approach is unsuitable for graphs containing blank nodes as BNode serialization *will* differ and afaik, the only mooted solution to this is Digital Bazaar's URDNA2015 RDF Dataset Canonicalization <https://json-ld.github.io/rdf-dataset-canonicalization/spec/> proposal, also now a topic in the RDFLib Github discussions section <https://github.com/RDFLib/rdflib/discussions/1545> -- http://github.com/RDFLib --- You received this message because you are subscribed to the Google Groups "rdflib-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to rdflib-dev+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/rdflib-dev/f582793a-e22d-404c-9e6d-f3d901a7e76cn%40googlegroups.com.
