Allow index's sid to be pulled in query results

[bplatz]

With json-ld, we completely hide the underlying index's subject id integers in query results.

There may still be some advanced circumstances where getting this is desired (working with one now). This re-enables the ability to request :_id in the subject crawl as a special keyword to request the underlying sid.

ex this query in the test

@(fluree/query db {:select {'?s [:_id :* {:ex/favArtist [:_id :schema/name]}]}
                   :where  [['?s :type :ex/User]]})

Produces this result:

[{:_id          211106232532999,
  :id           :ex/bob,
  :rdf/type     [:ex/User],
  :schema/name  "Bob",
  :ex/favArtist {:_id         211106232533000
                 :schema/name "Picasso"}}
 {:_id         211106232532998,
  :id          :ex/alice,
  :rdf/type    [:ex/User],
  :schema/name "Alice"}]

[dpetran]

Can you say a bit more about the use case?

[bplatz]

Comparing values across different query results are extremely fast when comparing integers, which is the whole reason our indexes use them to begin with.

Also for this reason dbproto/-subid is used extensively in our code.

This allows our nice graph crawl mechanism to expose the same benefits where you'd otherwise have to do a subsequent blocking async -subid call afterwards - which would not only slow things down but is redundant because this information was already known in the query results processing.

[zonotope]

📇