I gave up on sphinx and went to solr. I feel it is more mature. For example, sphinx didn't have an auto start init script and they tried to hit me up for consultancy fees cos I asked a simple question.
I use php and use solarium php client. Nice oop interface. Solr has a great community. My initial struggles were with getting it running, mostly because I don't know much about tomcat and it didn't just work for me as documented, but once i stumbled through it was ok. My search results accross 200k documents is instant on a small 512mb rackspacecloud instance so you will have no probs at all using solr for your needs. Sent from my iPhone On 21/02/2012, at 3:32 AM, Em <mailformailingli...@yahoo.de> wrote: > Hi James, > > I can not speak for Sphinx, since I never used it. > However, from reading your requirements there is nothing that fears Solr. > > Although Sphinx is written in C++, running Solr on top of a HotSpot JVM > gives you high performance. Furthermore the HotSpot JVM is optimizing > your code at runtime which sometimes allows long-running applications to > run as fast as software written in C++ (and sometimes even faster). > > Given that Solr is pretty fast and scalable (90k docs are a really small > index), you should have a closer look at the features each search-server > provides to you and how they suit your needs. > > You should always keep in mind that users will gladly wait a few > milliseconds longer for their highly-relevant search-results, but do not > care about a blazing fast 5ms response-time for a collection of > trash-results. > So try to find out what your concrete needs in terms of relevancy are > and which search-server provides you the tools to go. > I am pretty sure that both projects provide you php-client-libraries > etc. for indexing and searching (Solr does). > > Kind regards, > Em > > Am 20.02.2012 16:20, schrieb Spadez: >> I am creating what is effectively a search engine. Content is collected via >> spiders at >> then is inserted into my database and becomes searchable and filterable. >> >> I invision there being around 90K records to be searched at any one time. >> The content is >> blog posts and forum posts so we are basically looking at full text with >> some additional >> filters based on location, category and date posted. >> >> What is really important to me is speed and relevancy. The index size or >> index time >> really isn’t too big of an issue. From the benchmarks I have seen it looks >> like Sphinx >> is much faster at querying data and showing results, but that Solr has >> improved relevancy. >> >> My website is coded entirely in PHP and I am planning on using a MYSQL >> database. Can >> anyone please give me a bit of input and help me decide which product might >> be better >> suited to me. >> >> Regards, >> >> James >> >> -- >> View this message in context: >> http://lucene.472066.n3.nabble.com/Is-Sphinx-better-suited-to-me-or-should-I-look-at-Solr-tp3760988p3760988.html >> Sent from the Solr - User mailing list archive at Nabble.com. >>