The free Java search engine Apache Lucene is available in version 2.9. The release provides numerous new features and API enhancements.
Among the new features to Apache Lucene are per segment searching and more scalable multi-term queries. New analyzers are available for Persian, Arabic and Chinese. Also, developers increased the speed of numeric range searching so that they need not be converted to text and are indexed in a tree structure for faster processing.
Version 2.9 is not completely backward-compatible. Details are in the Lucene changelog. The team recommends compiling applications with Lucene 2.9 and not simply swap out the binaries. That way problems with backward compatibility can be ironed out more easily.
The next Lucene release should bear the number 3.0, require Java 1.5 and remove "deprecated" code.
Even state-of-the-art computers need to use clever methods to process ever-increasing amounts of document data. The open source Lucene framework uses inverted indexing for fast searches of document collections.