Can I be so bold as to ask what you were expecting too understand with such a large offset? You mention this is one of many tests. Are you performing a comparison of different implementations? 20M doesn't seem to be a useful offset. IMHO if you need to offset that far the query needs some work! Plus without some form of transaction control the underlying data might change between calls rendering the idea of paging mute. The results seem somewhat academic and whilst the different queries take different times the whole thing seems contrived.
Dick -------- Original message --------From: Laura Morales <laure...@mail.com> Date: 16/12/2017 15:58 (GMT+00:00) To: users@jena.apache.org Cc: users@jena.apache.org Subject: Re: Very very slow query when using a high OFFSET So, as a comparison this is the jconsole overview when using defaultGraph instead of namedGraph https://lut.im/5gKUNX0MdF/iRpiXLfTtDNapAIU.png The CPU spike at the start of the graph is when I issued the query, which returned before the CPU graph returns to zero. Unfortunately jconsole seems to sample data every few seconds, so I think the graph is not very accurate because the query returned more quickly than jconsole is sampling. But it's good enough to see the difference with the previous graph.
