- All Implemented Interfaces:
public class CachedFilterBuilder
- extends Object
- implements FilterBuilder
Filters are cached in an LRU Cache keyed on the contained query or filter object. Using this will
speed up overall performance for repeated uses of the same expensive query/filter. The sorts of
queries/filters likely to benefit from caching need not necessarily be complex - e.g. simple
TermQuerys with a large DF (document frequency) can be expensive on large indexes.
A good example of this might be a term query on a field with only 2 possible values -
"true" or "false". In a large index, querying or filtering on this field requires reading
millions of document ids from disk which can more usefully be cached as a filter bitset.
For Queries/Filters to be cached and reused the object must implement hashcode and
equals methods correctly so that duplicate queries/filters can be detected in the cache.
The CoreParser.maxNumCachedFilters property can be used to control the size of the LRU
Cache established during the construction of CoreParser instances.
|Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
public CachedFilterBuilder(QueryBuilderFactory queryFactory,
public org.apache.lucene.search.Filter getFilter(Element e)
- Specified by:
getFilter in interface
Copyright © 2000-2010 Apache Software Foundation. All Rights Reserved.