Output dataset section is missing

[dachafra]

The spec misses the section for the output dataset (see R2RML section).

[DylanVanAssche]

FYI: This section would need to be in sync with RML IO's Targets.
Within Target's there's a default target of a processor (up to the implementation to define its default target).

[dachafra]

Does this mean that the default output of the generation process is not a file?

[DylanVanAssche]

Currently no, the default is up to the system at the moment.

Imagine a stream processing RML based engine: the default might be a MQTT stream again as default output.
A RMLMapper-like engine will pick a file, an engine baked into a triplestore (to ingest non-RDF data directly into the store) would have the internal triplestore data structure as default output.

[dachafra]

Could you point to the part of the IO spec where it is described?

[DylanVanAssche]

https://w3id.org/rml/io/spec#logical-target-in-rml

[dachafra]

The output-dataset does not need to be accommodated to the IO, because this is RML-Core, RML-IO at some point should provide its own algorithm.

This is fixed in dachafra@e99d686, but we need to wait for the approval of #244 to make the PR