Hi Kai No, I have no problem returning hits.
When I do have problems like this, I usually find I have something more to learn about Lucene indexing. Try looking at the data and query in Luke. I usually find this is the best way to understand what is going on. Here is the link to Luke: http://www.getopt.org/luke/ I hope you get it sorted. Patrick On 07/08/07, Kai Hu <[EMAIL PROTECTED]> wrote: > By the way, Patrick,did you have a problem that IndexSearcher.search(Query > query) cann't get the all matched hits.it only return a part of matched hits. > my test code is: > String key = "title:good"; > Directory directory = FSDirectory.getDirectory("d:\\index\\"); > IndexSearcher searcher = new IndexSearcher(directory); > QueryParser queryParser = new QueryParser("",analyzer); > Query query = queryParser.parse(key); > hits = searcher.search(query,sort); > there are two documents indexed which "title" value is "good",but when I > searched by key"title:good",it returned only one documents.is it a bug? > > kai > > >Hi Kai > >We keep a synchronized map of LuceneIndexAccessor instances, one instance per > >The map is keyed on the directory path. We then re-use > >the accessor rather than creating a new one each time. > > > >Patrick > On 06/08/07, Kai Hu <[EMAIL PROTECTED]> wrote: > > Thanks , Patrick, > > > > It is useful. But I found a problem that I use > > > > "new LuceneIndexAccessor(accessProvider);" when a request comes in B/S,the > > LuceneIndexAccessor.getWriter() will lose its sense,it will new an > > IndexWriter. > > > > public IndexWriter getWriter() throws IOException { > > > > IndexWriter result; > > > > synchronized (this) {//here "synchronized" will lose its sense > > > > checkClosed(); > > > > ... > > > > if (cachedWriter != null) { > > > > log.debug("returning cached writer"); > > > > result = cachedWriter; > > > > writerUseCount++; > > > > } else { > > > > log.debug("opening new writer and caching it"); > > > > result = accessProvider.getWriter();// when new a > > LuceneIndexAccessor Object ,it will new an IndexWriter Object > > > > cachedWriter = result; > > > > writerUseCount = 1; > > > > } > > > > } > > > > } > > > > It will also throw a Exception "cann't obtain the Lock",should I use a > > single instance of LuceneIndexAccessor? Suppose I use a single instance of > > LuceneIndexAccessor,how to set a different Directory or Analyzer at one > > time. > > > > > > > > kai > > > > > > /////////////////////////////////////////////////////////// > > > > /////////////////////////////////////////////////////////// > > > > > > > > >Hi Kai > > > > > > > > > >We use the Lucene Index Accessor contribution: > > > > > > > > > >http://www.nabble.com/Fwd%3A-Contribution%3A-LuceneIndexAccessor-t17416.html#a47049 > > > > > > > > > >Patrick > > > > > > > > On 06/08/07, Kai Hu <[EMAIL PROTECTED]> wrote: > > > > > Hi, > > > > > > > > > > How do you solve the problems when add,update,delete documents > > > > > in muti-threads,use synchronized ? > > > > > > > > > > > > > > > > > > --------------------------------------------------------------------- > > > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
