15 Dec, 2008
My Lucene.Net experience - Part 2
Posted by: Carl In: asp.net| c#| development| lucene.net
Since my first post, I’ve been tinkering a bit more with Lucene.Net. I’ve added a few more columns to index, determined that a SimpleAnalyzer would fit my needs by tokenizing every word by not removing “stop words”, and I’ve created a Windows Service to do a multi-threaded index of my 350,106 rows of data. I’ve definitely got some bottlenecks in my code, partially due to the logging that I’m doing, as it’s a shared resource, and each thread has to stop and wait to get access in order to log. Future versions of my indexer will use separate log files per thread, and merge them via another thread. At least, that’s my hope. I haven’t tried log4net yet, and as such, I’m not sure if it will alieviate these issues, but I imagine they will since it requires less “reinventing the wheel” on my part.
I’m still digging this as an alternative to SQL Server Full-Text Indexing, as it just would not get me the results I desired, with multiple refinements and countless Google-searching. I’m curious to hear back from those using SQL Server 2008, and it’s integrated indexing services, but, for now, Lucene.Net is a good here-and-now solution that I will stick with.
Just thought I’d get that out there. That is all :p