Hi list,
Let me describe my issue taking a simpler model. Lets say I were to build a
blog which allows each post to have multiple keywords. I want to provide a
search over the posts but restricted to a subset of the keywords (say -
python, windows, etc.). How can I structure the index in this c
index is locked when IndexReader is opening index. After that (when
index structure is read) lock is released.
Aigner, Thomas wrote:
Howdy all,
I am running some tests against our search. I have created a
script that will run 30 simultaneous executions of 25 searches each.
The 25 sea
Howdy all,
I am running some tests against our search. I have created a
script that will run 30 simultaneous executions of 25 searches each.
The 25 searches are randomized but the same data is used. I am
consistently getting "Lock obtain timed out:
Lock@/tmp/lucene-77053e9cdda0a530f729e7
The StandardAnalyzer should work just fine with it, It will break the
search string to 5 search terms.
HTH
Aviran
http://www.aviransplace.com
_
From: Karthik N S [mailto:[EMAIL PROTECTED]
Sent: Friday, August 05, 2005 1:57 AM
To: LUCENE
Subject: Split Search Word
Hi Luceners
Apol
You can try working with two indexes one for all of today's messages
which will be pretty small, and another for past messages. Then once a
day merge the small index to the big one and start fresh. This way you
need only to open an IndexReader for the small index while the big one
does not change.
You might want to look at stemming for "de pluralization" it boils down words
to their "root"
So bombs and bomming get stemmed to bomb.
I'm using the snowball stemmer, which handles different languages as well as
english.
It is in the sandbox.
org.apache.lucene.analysis.snowball.SnowballFilt
Mufaddal Khumri wrote:
Are there
analyzers that do this already?
Its not an analyzer, but the "norm" feature of this tool does a good job
at getting to the normalized form of the words...
http://umlslex.nlm.nih.gov/lvg/current/
http://umlslex.nlm.nih.gov/lvg/current/docs/userDoc/norm.htm
Hi Martin!
It's starts to work know, thanks !! :-)
>In my case "text" is everything. Whenever I add anything to the index, I
>also add a "text" Field. So in your case, I would build the index like
>you did with an additional
>doc.add(Field.Text("text", bean.getCompanyName()));
I see, but I do
It seems it doesn't change the boost factors and that is strange, I
did as in the manual, I thought it was easy, I just use:
fieldname.setBoost(1.2f);
and then I add to the doc that field, and I index the doc.
All the indexing process goes ok, only these boost factors are
different from what I do
