Sphinx full text search
January 11, 2007
During the mysql meeting last night, Tobias mentioned the sphinx search engine project. Apparently some ex-mysql guys split off and formed a consulting company, and are also working on this project. I don’t know the nitty-gritty details, but I do know this seems pretty fast. Tobias said it smokes Lucene for speed, and my initial test seemed to agree. I just did one test that indexed 600,000 records - each record had probably 2k of text max, so it’s not huge - in 37 seconds. This is running under vmware on a machine with other vmware servers running on it - I don’t have the specs offhand, but it’s not a top of the line machine. Queries are all coming back in under .010 seconds, matching 2000+ results. I will run some more tests to see how it scales and post more info here if there’s anything more notable. Definitely looks like a project to watch.
Did you like this post? Buy me a hot chocolate!
Posted in Database, Development, MySQL, Software, Uncategorized
