java - lucene index lookup query slower than where clause based query for more data -
i facing issues lucene index lookup in start clause slower clause. e.g
i running neo4j 1.9.m05 on mac java process having 4 gb of heap 1.5 milllion nodes , 5 million relationships , 5 million properties.
the query looks below....basically has match , other conditions. trying optimize lot of queries index lookup, changing existing query not have index lookup, looks index lookup slower
cypher 1.8 start a=node(21102), cu=node:myindex(id="020513_1359970281348_1360024480338_2") match cu<-[:cp_in]-(cp), (out)<-[:ap]-(cp), (out)-[:ac]->(path0_child)-[:acp*0..10]->(path0) ,(out)<-[:ahp]-(a) path0.id! =~'(?i)^\\qtest\\e$' return distinct out skip 0 limit 50 we using spring data , putting index lookup cache not seem help
index<node> index = template.getgraphdatabase().getindex("myindex"); ((luceneindex<node>) index).setcachecapacity("id", 30000000); any clue or advice appreciated
regards
sai
Comments
Post a Comment