Load data in memory

[adel-ayara]

Hello,

Is Qlever in-memory triple store or it's a SPRQL engine that has to be connected to external triple store?
Is it possible to load and index the data to be queried directly in-memory ?

Best regards,
Adel

[hannahbast]

@adel-ayara QLever stores its index data structures on disk, preferably on NVMe SSD. But if you have enough RAM, then all the data structures will eventually be cached in RAM. So you get both, depending on what hardware you have. Try to build an index yourself following the instructions under https://github.com/ad-freiburg/qlever-control

[adel-ayara]

Thanks @hannahbast again for your feedback!
by get both, do you mean that is there a parameter to define how and where does Qlever the sata: in memory or on disk?
Another question: what is the difference between the get-data and index commands?

[hannahbast]

The get-data command is just a shortcut to download the data. For example, for Wikidata, the comman downloads the latest RDF dump from https://dumps.wikimedia.org/wikidatawiki/entities.

The index command precomputes index data structures so that QLever can process queries on the data quickly. Every SPARQL engine does that. The engines differ in exactly which data structures they precompute and how efficient the query processing is.

The index data structures are written on disk. But if you have enough RAM, the operating system will eventually cache them in RAM.

Do you have something like a disk in your setting?

[adel-ayara]

Yes we have a disk but its access is very complicated and can't be done via classical system routines since it's a safe embedded system