Currently triples are read from the database without any order specified. This is fast but leads to ugly RDF output, especially for text/turtle and application/rdf+xml. We can either request ordering on the SQL query level or follow the application/ld+json kind of approach where we cache triples up to some size and if they fit the buffer, we order them.
Currently triples are read from the database without any order specified. This is fast but leads to ugly RDF output, especially for text/turtle and application/rdf+xml. We can either request ordering on the SQL query level or follow the application/ld+json kind of approach where we cache triples up to some size and if they fit the buffer, we order them.