GitHub user arne-bdt created a discussion: Why is Quad not a composition of the
graph Node + Triple?
Working with Datasets, we constantly create quads and triples converting one
into the other.
This costs a lot of performance and memory.
I have several ideas to address this:
1. Make `Quad` a composition of `Node` + `Triple`
--> a lot of code could be optimized this way
2. Make `Quad` an interface with two Implementations `QuadOfNodes` and
`QuadOfTriple`
--> this way, we could decide in each implementation, which one fits best.
Where Nodes are stored, use `QuadOfNodes`. Where Triples are stored,
use `QuadOfTriples`.
3. If the upper options are not okay, Extend the
`org.apache.jena.sparql.core.DatasetGraph` with `(Node g, Triple t)` overloads.
This way, one can use the default implementation or implement it directly
where this is beneficial.
4. If `Triple` was an interface too, a `Quad` could also extend that
interface. This way, we could even store quads in graphs directly and would in
many cases not have to convert them again, when used in a `DatasetGraph`
What do you think?
GitHub link: https://github.com/apache/jena/discussions/3967
----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
