Apache Lucene 8.2.0 released, Lucene is a full-text search engine architecture that provides a complete query engine and index engine, part of the text analysis engine. The purpose is to provide a simple-to-use software developer kit for convenient full-text search functions in the target system, or as a basis to establish a complete full-text search engine.
The new version of the highlights include:
API Changes
- Query interval to move from the sandbox to the query module
New features
- For indexing and query new XYShape general Cartesian geometry of fields and queries
- Snowball interference for Estonian / analyzer
- FeatureSortfield provided to allow the value of the function in descending order of search hits Sort
- Add a new KoreanNumberFilter, Korean characters can be changed to digital processing and decimal point
- Adding support for doc-value range field
- Add Monitor subproject (previous Luwak monitor library), allowing an efficient manner the document flow with a set of queries registered match
- Add a query using the index value range ordering in a sandbox
optimization
- Use the search index in IntArrayDocIdSet # advance method rather than binary search
- If IndexSearcher The executive, passing a thread of execution is used. At least now calling thread to perform a search index, even if the execution of the program provided to minimize thread context switching
- The new storage policy BKD tree has a low base, can reduce storage costs, you can use when searching to speed up queries
- In BlockDocsEnum and BlockImpactsEverythingEnum needed only if lazy loading