Hi, Usually, limit=-1 works as a single pass-through and counts accumulating; but when limit >0 causes collecting per value docset, whic might take longer. There's a note about this effect in uniqueBlock() description.
On Wed, Jul 8, 2020 at 11:29 AM ana <anasolr3...@gmail.com> wrote: > Hi Team, > Which is more optimized: facet.limit=-1 OR facet.limit=10000/50000/40000? > For a high Cardinality string field, with no cache enabled, no docValues > enabled, after every RELOAD on Solr admin UI for each query with different > facet.limit, why the QTime for "facet.limit=-1" is lesser as compared to > that of a 'facet.limit=5000/10000". What factors apart from those listed > above matters in calculating QTime? > > My understanding is that facet.limit=-1 should have higher response Time as > per Solr ref guide as compared to any other higher facet.limit specified. > > Experiment : > > field = abc_s > cardinality:71520 > num of docs : count:52055449, > total num of facets:70657 > appliedMethod: FC > Test query : > http://localhost:8983/solr/ > <collection_name>/select?facet.field=abc_s&facet=on&q=*:*&rows=0&debug=true&f.abc_s.facet.limit=-1 > > facet.limit -1 100 5000 10000 40000 > 50000 > QTime 983 857 3429 5324 1006 > 1027 > > Debug response for facet.limit=10000 is attached > facet_Response_10000.txt > <https://lucene.472066.n3.nabble.com/file/t495711/facet_Response_10000.txt> > > > > > -- > Sent from: https://lucene.472066.n3.nabble.com/Solr-User-f472068.html > -- Sincerely yours Mikhail Khludnev