I use Lucene/MemoryIndex for a large number of queries against data in a streaming system. I'm looking to upgrade from v3.5 to 4.x, but it seems that using MemoryIndex is roughly 25% slower based on a benchmark I built using our internal queries and a sample of 1000 documents to run against. I haven't seen any benchmarks in the mailing list or anyone talking about a slowdown and am wondering if the classes used in 3.x are contributing to what I see. Are there any known classes that should not be used with Lucene 4.x? Perhaps a different QueryParser to make more efficient queries to run against the MemoryIndex, or Analyzers that shouldn't be used?
If the MemoryIndex *is* slower now, would moving to a RAMDirectory and batching queries be a better choice? I've been using MemoryIndex based on the javadoc description for some time, but need to see better performance before moving up to 4.x. Thanks, Chris -- View this message in context: http://lucene.472066.n3.nabble.com/MemoryIndex-in-Lucene-4-x-tp4077919.html Sent from the Lucene - Java Users mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org