Marklogic : advantage of using embedded triples over triple store

Question

I have two questions related to embedded triples.

1. Can we apply inferencing through the use of rulesets?
2. Can we do a selective updates on triples given a subject and predicate?

Essentially, the bigger question is if the inferencing and updates on embedded triples are available, what would be the motivation to use triple store?

Answer 1

1) Yes, triples are all the same, regardless how they are injected in MarkLogic.

2) If the triples are embedded in XML or JSON, you can modify them by modifying the XML or the JSON (using the usual update functions/mechanisms). SPARQL Update can be used only for managed triples.

Essentially, the bigger question is if the inferencing and updates on embedded triples are available, what would be the motivation to use triple store?

As soon as you use triples, you use the triple store. You can call it the triple engine if you prefer. Regardless the way they are created in MarkLogic, all triples, once identified as such, are "stored" in the triple index.

Using triples is using the triple store.