On Friday, November 21, 2003, at 02:34 PM, Jianshuo Niu wrote:
I read your post on lucene bug list. However, I try the change you
suggested, but it just changed "t-shirts" to "shirt".
What Analyzer are you using?
-
To unsubscrib
Dear Victor:
I read your post on lucene bug list. However, I try the change you
suggested, but it just changed "t-shirts" to "shirt".
I downloaded lucene1.3-rc1 source, changed the above line in
QueryParser.jj, and recompiled the source. After the change, the query I
got is:
+(name:shirt)
be
You are right, just for search it wont take much time. I have relatively
small index.
Almost all documents change daily. So one way I am thinking of updating
index is creating new index set in a temp directory by other process. When
time comes to update index then renaming in-use dir to archiv
There are some important questions to ask before you discount performing all
of those searches. What are your performance requirements? How big is your
index? How are you deploying it?
-Original Message-
From: sam s [mailto:[EMAIL PROTECTED]
Sent: Thursday, November 20, 2003 8:15 PM
To:
for best results, you want to suggest from the dictionary if indexed terms in the
collection. if the index has been created with named entities (not impossible but hard
in the current Lucene system) you can get it pretty exact.
Herb...
---
If you are going to be searching the same index, just keep that
instance of IndexSearcher around, and close it only when you no longer
need it, or when you detect that the index changed (e.g. docs
added/deleted), at which point you will want to close the old
IndexSearcher (or just throw it away and