fixed

issue #438

imixs-workflow-engine/src/main/java/org/imixs/workflow/engine/lucene/LuceneSearchService.java

@@ -218,9 +218,10 @@ public List<ItemCollection> search(String sSearchTerm, int pageSize, int pageInd
 			parser.setAllowLeadingWildcard(true);
 
 			// set default operator?
-			if (defaultOperator != null)
+			if (defaultOperator != null) {
 				parser.setDefaultOperator(defaultOperator);
-
+			}
+
 			long lsearchtime = System.currentTimeMillis();
 			TopDocs topDocs = null;
 			TopDocsCollector<?> collector = null;
@@ -244,7 +245,7 @@ public List<ItemCollection> search(String sSearchTerm, int pageSize, int pageInd
 				// sorted by sortoder
 				logger.finest("......lucene result sorted by sortOrder= '" + sortOrder + "' ");
 				// MAX_SEARCH_RESULT is limiting the total number of hits
-				collector = TopFieldCollector.create(sortOrder, maxSearchResult, false, false, false);
+				collector = TopFieldCollector.create(sortOrder, maxSearchResult, false, false, false, false);
 
 			} else {
 				// sorted by score
@@ -386,12 +387,7 @@ public int getTotalHits(String sSearchTerm, int maxResult, Operator defaultOpera
 	 * extended with a users roles to test the read access level of each workitem
 	 * matching the search term.
 	 * 
-	 * The optional param 'defaultOperator' can be set to Operator.AND
-	 * 
 	 * @param sSearchTerm
-	 * @param defaultOperator
-	 *            - optional to change the default search operator
-	 * 
 	 * @return extended search term
 	 * @throws QueryException
 	 *             in case the searchtem is not understandable.
@@ -504,6 +500,12 @@ QueryParser createQueryParser(Properties prop) {
 			logger.finest("......DefaultOperator: AND");
 			parser.setDefaultOperator(Operator.AND);
 		}
+
+		// set setSplitOnWhitespace (issue #438)
+		String splitOnWhitespace = prop.getProperty("lucene.splitOnWhitespace","true");
+		boolean bSplitOnWhitespace=Boolean.parseBoolean(splitOnWhitespace);
+		logger.finest("......SplitOnWhitespace: "+bSplitOnWhitespace);
+		parser.setSplitOnWhitespace(bSplitOnWhitespace);
 
 		return parser;
 	}

src/site/markdown/engine/luceneservice.md

@@ -33,7 +33,9 @@ The custom configuration of the _LuceneService_ can be provided in the file _imi
 	# Fields to be added into the searchindex
 	lucence.fulltextFieldList=txtsearchstring,txtSubject,txtname,txtEmail,txtWorkflowAbstract,txtWorkflowSummary
 	lucence.indexFieldListAnalyze=
-	lucence.indexFieldListNoAnalyze=datDate,txtWorkflowGroup,txtemail, datdate, datfrom, datto, numsequencenumber, txtUsername,
+	lucence.indexFieldListNoAnalyze=datDate,txtWorkflowGroup,txtemail, datdate, datfrom, datto, numsequencenumber, txtUsername
+	lucene.defaultOperator=AND
+	lucene.splitOnWhitespace=true	
 
 
 ### IndexDir
@@ -50,8 +52,13 @@ The property 'lucene.indexFieldListAnalyze' defines a comma separated list of fi
 ### IndexFieldListNoAnalyze
 The property 'lucene.indexFieldListNoAnalyze' defines a comma separated list of fields which will be added as keyword  fields into the lucene index. The content of this fields will not be analyzed. So a exact phrase search is possible here.
 
+### defaultOperator
+The defaultOperator sets the boolean operator of the QueryParser. In default mode (AND\_OPERATOR) terms without any modifiers are considered optional: for example _capital of France_ is equal to _capital AND of AND France_.
+In OR\_OPERATOR mode terms are considered to be in conjunction: the above mentioned query is parsed as _capital OR of OR France_
+
+### splitOnWhitespace
+Whether query text should be split on whitespace prior to analysis. Default is true. For example _cat dog_ will be treated as _cat AND dog_
 
-
 ## How to Initialize the Lucene Index
 
 The lucene index is automatically written into the Index Directory by the Imixs-Workflow engine.