This might be related to filesystem, internal lucene buffering/caching,
or practically anything that an implementor does not need to have
knowledge of.
The only thing that you, the implementor, *do* need to know is that you
should *not* access a Hits object after the searcher is closed ;)
/Ronnie
Huinan wrote:
Thanks, Ronnie. But why it works in some cases (when there is a small
number
of documents inside the index) ?
On 9/13/06, Ronnie Kolehmainen <[EMAIL PROTECTED]> wrote:
Do not close the searcher until you are done with the Hits object.
See the javadocs for Searchable.close()
http://lucene.apache.org/java/docs/api/org/apache/lucene/search/Searchable.html#close()
/Ronnie
Huinan wrote:
> Hi,
>
> I'm having a weird problem:
>
> I created an index using IndexWriter. Then I had a piece of code which
> searches the index, then print out a particular field of the first
document
> of the hits.(See the following code) As simple as that.
>
> Hits hits = IndexSearchUtil.getHits(defaultIndexLocation, "A",
"a");
> System.out.println(hits.length()); // This prints 1.
> Document doc = hits.doc(0); // <------But this will throw
an
> IOException, and this is the problem.
> System.out.println(doc.get("A"));
>
> The strange thing is that it DOES NOT ALWAYS throw an IOException. When
I
> have a small number (tens) of index entries, this works fine. But
beyond
a
> certain threshold, It begins to throw up. :-(
>
> Does anyone have the same problem? or could suggest what might have
gone
> wrong?
>
> Thanks a lot!
>
>
> Regards.
>
> Huinan
>
>
> Appendix
> (inside IndexSearchUtil class)
> public static Hits getHits(String indexLocation, String fieldName,
> String key) throws IOException {
> return getHits(indexLocation, fieldName, key, true);
> }
>
> public static Hits getHits(String indexLocation, String fieldName,
> String key, boolean fieldTokenized) throws IOException {
> IndexSearcher searcher = new IndexSearcher(indexLocation);
> if (!fieldTokenized)
> key = "\"" + key + "\"";
> QueryParser parser = new QueryParser(fieldName, new
> KeywordAnalyzer());
> Query query;
> try {
> query = parser.parse(key);
> Hits hits = searcher.search(query);
> searcher.close();
> return hits;
> } catch (ParseException e) {
> e.printStackTrace();
> return null;
> }
> }
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
--
________________________________________
Ronnie Kolehmainen
Systems Developer
Electronic Publishing Centre
Uppsala University Library
+46 (0)18 471 5847
mailto:[EMAIL PROTECTED]
________________________________________
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
