Named Query support

[WolfgangFahl]

Would be great if you take part in the discussion see WDscholia/scholia#2412

[hannahbast]

@WolfgangFahl Is there a concrete question? Besides, could you update https://wiki.bitplan.com/index.php/Get_your_own_copy_of_WikiData . QLever can now build an index for the complete Wikidata in 4.5 hours, requiring no more than 32 GB of RAM (AMD Ryzen 9, 16 cores).

[WolfgangFahl]

It would be great to get some endorsement and comments and to reconsider #859
I am adding a line to the success table - how many billion triples do we have at this time?

[hannahbast]

@Wikidata number of triples: You can always see this for yourself by going to https://qlever.cs.uni-freiburg.de/wikidata and clicking on "Index Information". The latest version of the dataset has 19.3 billion triples.

@your query from #859: This query has many issues that have nothing to do with QLever. First, there is SERVICE wikibase:label in it, which is specific for the WDQS. Second, there are several SPARQL mistakes. Third, even when I fix those, the query returns to many entities (over 90 million) because it does not restrict to proceedings at any point. You can see this yourself by clicking on the "Analysis" button (which also works if the query did not run to completion).

Here is the almost working query: https://qlever.cs.uni-freiburg.de/wikidata/tjtB9m . If you add the constraint to it that ?proceedings is actually of type proceedings, it will work.

[WolfgangFahl]

@hannahbast #859 is about almost 300 different queries - not a single one