Hi Curtis 3 sec without FullText index ? select * from Abstract where appln_abstract CONTAINSTEXT "of a pipe of the pipe"
can you post the explain of the previous query? How do you run the LUCENE query? 2014-08-19 10:50 GMT+02:00 'Curtis Mosters' via OrientDB < [email protected]>: > Wow, I tested > > select * from Abstract where appln_abstract CONTAINSTEXT "of a pipe of > the pipe" > > on 280k entries without index. It was running 3 sec till I got a result. > > Then I tested again on the Lucene indexed 280k database and it took 20 sec. > > So how can that be. I read that OrientDB is already indexing. But from > Neo4j I know that Lucene is much faster. But why in my case it's 7x slower? > Which indexer is used when you don't explicitly set an indexer? > > Am Montag, 18. August 2014 21:33:30 UTC+2 schrieb Enrico Risa: >> >> Hi Curtis >> >> the LIKE operator doesn't use the FULLTEXT index. >> >> Could you retry the query with the CONTAINSTEXT >> operator. It should be faster because rely on the FULLTEXT index >> >> http://www.orientechnologies.com/docs/1.7.8/orientdb.wiki/SQL-Where.html >> see here >> >> Enrico >> >> >> 2014-08-18 20:47 GMT+02:00 'Curtis Mosters' via OrientDB < >> [email protected]>: >> >> I now tried it with Lucene and the index creating is much *faster*. =) >>> >>> Also tested again both ways: >>> >>> 1. Importing without index: 120 sec + Indexing 80 sec >>> 2. Importing with index: 340 sec >>> + extracted 274.139 records (686 records/sec) - 274.139 records -> >>> loaded 274.13 >>> 8 vertices (686 vertices/sec) Total time: 339809ms [0 warnings, 0 >>> errors] >>> >>> >>> So is Lucene actually faster when building up the index afterwards? Or >>> is my computer really that crappy so that my 100% cpu usage really harming >>> the benchmark? >>> >>> They query from above was done in ~25 sec, so it's also a bit faster. >>> Can that be true? >>> >>> >>> Am Montag, 18. August 2014 17:23:40 UTC+2 schrieb Enrico Risa: >>> >>>> Hi Curtis >>>> >>>> can you post the result of >>>> >>>> explain select * from Abstract where appln_abstract LIKE "%of a pipe >>>> of the pipe%" >>>> >>>> Thanks >>>> Enrico >>>> >>>> >>>> 2014-08-18 17:19 GMT+02:00 'Curtis Mosters' via OrientDB < >>>> [email protected]>: >>>> >>>>> I'm still testing around with OrientDB. Today I realized that OrientDB >>>>> is 3 times slower on the same data, with the same indexer compared to >>>>> MySQL. How can that be? >>>>> >>>>> So there are ~250k entries. FULLTEXT indexer are used on both db's. >>>>> (from https://github.com/orientechnologies/orientdb/wiki/Indexes) >>>>> >>>>> And the test query is: >>>>> select * from Abstract where appln_abstract LIKE "%of a pipe of the >>>>> pipe%" >>>>> >>>>> in OrientDB: 34 sec >>>>> in MySQL: 14 sec >>>>> >>>>> I tested this on them both 3 times and this is the average. >>>>> >>>>> Any ideas? >>>>> >>>>> -- >>>>> >>>>> --- >>>>> You received this message because you are subscribed to the Google >>>>> Groups "OrientDB" group. >>>>> To unsubscribe from this group and stop receiving emails from it, send >>>>> an email to [email protected]. >>>>> >>>>> For more options, visit https://groups.google.com/d/optout. >>>>> >>>> >>>> -- >>> >>> --- >>> You received this message because you are subscribed to the Google >>> Groups "OrientDB" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> -- > > --- > You received this message because you are subscribed to the Google Groups > "OrientDB" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- --- You received this message because you are subscribed to the Google Groups "OrientDB" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
