lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chris Mannion" <chris.mann...@nonstopgov.com>
Subject No hits for longer search strings
Date Thu, 16 Oct 2008 10:14:34 GMT
Hi All

I have a bit of a puzzle in the Lucene system we've been running.  Part of
our use involves inserting documents indexed by a unique key and then
running exact searches to find that single document again later to display
(the documents are also indexed by several other fields and used in a
broader search, hence using Lucene).  I've just hit a strange bug lately
where the documents are not being found under certain circumstances.  We've
usually been indexing by unique ids about 12 characters in length, e.g.
t10001c53421 and this has always worked fine, however when the id length is
whacked up to 18, e.g. t10001c63241985103, the search never finds the
documents.  I've used Luke which allows me to open the index for browsing
and can see that the documents are definitely in the index and have the
correct unique id values stored in there, so it seems to be a problem with
the search.

We're using a search string like (unique_id:(t10001c63241985103*) ) and
standard analyzers to run the search.  Does anyone have any ideas why the
increased search string length would affect things like this?

-- 
Chris Mannion
iCasework and LocalAlert implementation team
0208 144 4416

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message