Tried to allocate 278 TB, ...

[WolfgangFahl]

Tried to allocate 278 TB, but only 28.9 GB were available. Clear the cache or allow more memory for QLever during startup

http://qlever.wikidata.dbis.rwth-aachen.de/wikidata/C50ESd

The error message seems misleading i doubt that 278 TB of memory need to be allocated for this query.

https://qlever.cs.uni-freiburg.de/wikidata/yAMX5c

runs fine

[tuukka]

Tried to allocate 278 TB, but only 28.9 GB were available. Clear the cache or allow more memory for QLever during startup

http://qlever.wikidata.dbis.rwth-aachen.de/wikidata/C50ESd

Just an observation that you had a typo in the filter, and the query works with the typo fixed: http://qlever.wikidata.dbis.rwth-aachen.de/wikidata/vKWePT

[WolfgangFahl]

@tuukka - this is about the error message not the cause of the error

[hannahbast]

@WolfgangFahl Again, I get a 404 for your links. But from the discussion (thanks, @tuukka), I can guess what the problem is.

Technically, the error message is correct. Consider the following simple example. Processing this query correctly requires computing a huge cross-product. I agree that would be nice to warn the user that maybe they made a mistake when formulating the query. But it's not the primary task of a SPARQL endpoint and it can be quite hard, depending on the mistakes.

SELECT * WHERE {
  ?person wdt:P31 wd:Q5 .
  ?pesron wdt:P106 ?occupation .
}

Try it on QLever

[hannahbast]

The above being said, a solution for this particular case might be to find out whether the query involves a cross-product (easy) and in that case add to the technically correct error message the following in parentheses:

(the processing of this query involves a cross-product, check if this is intentional or if you have mistyped a variable name)