Examples of org.apache.xindice.core.indexer.LuceneIndexer

org.apache.xindice.core.indexer.LuceneIndexer

LuceneIndexer is used for maintaining full text indexes. It operates on documents instead of elements and allows to search for documents using native Lucene query. There can be only one LuceneIndexer per collection, however, it may have more than one IndexPattern.

Every IndexPattern corresponds to a Lucene document field. For every Xindice document, value of all matching elements will be indexed by a single Lucene document, allowing to search across the patterns.

Sample LuceneIndexer configuration:

 <index name='fulltext' class='org.apache.xindice.core.indexer.LuceneIndexer' analyzer='org.apache.lucene.analysis.SimpleAnalyzer'> <pattern pattern='meta@title' alias='title'/> <pattern pattern='description' alias='text'/> </index>

To search over this sample index, one could issue a query "title:tutorial AND text:xml".

For more details about LuceneIndexer configuration please see documentation for {@link #setConfig(org.apache.xindice.util.Configuration)}

@author Andy Armstrong @version $Revision: 586647 $, $Date: 2007-10-19 20:32:43 -0400 (Fri, 19 Oct 2007) $

            IndexPattern pattern = new IndexPattern(symbols, ps, nsMap);


            // check if there is full text indexer for this collection
            Indexer idx = context.getIndexManager().getBestIndexer(Indexer.STYLE_FULLTEXT, pattern);
            if (idx instanceof LuceneIndexer) {
                LuceneIndexer textInd = ((LuceneIndexer) idx);
                analyzer = textInd.getAnalyzer();
                parameters.put(PARAM_ANALYZER, analyzer);


                // see if index has matching pattern
                String alias = textInd.getPatternAlias(pattern);


                if (alias != null) {
                    // Queries that contain 'NOT', '!', '-' operators cannot be used here
                    // because LuceneIndexer searches for documents, in that context
                    // "NOT term" query means to find documents where 'term' does not
                    // appear in certain field at all. For XPath, however, it means that
                    // 'term' must not appear in text of the element that currently under
                    // evaluation, but may appear in the other elements that match the
                    // same IndexPattern.
                    //
                    // To make sure that all potentially matching documents are returned
                    // by the search, all subqueries with these operators are ignored
                    // on this step.
                    org.apache.lucene.search.Query parsedQuery = new SpecialQueryParser(alias, analyzer).parse(query);
                    IndexMatch[] matches = textInd.queryMatches(parsedQuery);
                    Key[] keys = QueryEngine.getUniqueKeys(matches);


                    return new NamedKeys(ps, ps.indexOf('@') != -1, keys);
                }
            } else {

            String alias = (String) i.next();
            config += "<pattern pattern='" + patterns.get(alias) + "' alias='" + alias + "' />";
        }
        config += "</index>";


        LuceneIndexer ind = (LuceneIndexer) collection.createIndexer(new Configuration(DOMParser.toDocument(config)));
        Thread.sleep(2000);
        return ind;
    }

Examples of org.apache.xindice.core.indexer.LuceneIndexer

Related Classes of org.apache.xindice.core.indexer.LuceneIndexer