Hi Jose,
Result set limits are capped at 1000 result rows.
If you are using Java, you could e.g. use our "jena-sparql-api"
project (
https://github.com/AKSW/jena-sparql-api), which provides
classes that enable you to work around the result set size limit
pretty easily (this library also supports caching each of the
individual retrieved result sets, such that if you have to restart
your process, it will run on cache up to the point where it was
cancelled).
Can you please give a concrete example of where you think something
goes wrong? The easiest would be to post the full HTTP urls you use
to retrieve the SPARQL result sets.
Cheers,
Claus