On Tuesday 05 December 2006 03:49, Zhang, Lisheng wrote:
I found that search time is about linear: 2nd time is about 2 times
longer than 1st query.
What exactly did you measure, only the search() or also opening the
IndexSearcher? The later depends on index size, thus you shouldn't re-open
Hi,
In my test case, four Quartz jobs are starting each third minute storing
records in a database followed by an index update.
After doing a test run over a period of 16 hours, I got this exception
after 10 hours:
java.io.IOException: Access is denied
at
Forgot something...
Also I got this exception, which may be related:
java.io.IOException: Cannot delete C:\dknewscenter\2\_5d.cfs
at
org.apache.lucene.store.FSDirectory.create(FSDirectory.java:319)
at
org.apache.lucene.store.FSDirectory.getDirectory(FSDirectory.java:208)
Hi,
thank you both for your help. Where would I find this Contributions?
Aaron
Risov, Maria wrote:
It's in Contributions rather than being in the core Lucene folder.
Marie Risov
-Original Message-
From: Erick Erickson [mailto:[EMAIL PROTECTED]
Sent: Monday, December
Zhang, Lisheng wrote:
Hi,
I indexed first 220,000, all with a special keyword, I did a simple
query and only fetched 5 docs, with Hits.length()=220,000.
Then I indexed 440,000 docs, with the same keyword, query it
again and fetched a few docs, with Hits.length(0=440,000.
I found that search
[EMAIL PROTECTED] wrote:
Forgot something...
Also I got this exception, which may be related:
java.io.IOException: Cannot delete C:\dknewscenter\2\_5d.cfs
at
org.apache.lucene.store.FSDirectory.create(FSDirectory.java:319)
at
[EMAIL PROTECTED] wrote:
Hi,
In my test case, four Quartz jobs are starting each third minute storing
records in a database followed by an index update.
After doing a test run over a period of 16 hours, I got this exception
after 10 hours:
java.io.IOException: Access is denied
at
Thank you for quick and detailed answer.
In this system multiple threads will, occasionally, try to write and/ or
read the same index, hence the pause waiting for the lock. This is not a
good way to implement it and was done as a temp solution for debug
purposes only. Multiple processes may
Hi,
I'm building an application that's going to classify some documents. So i have
a set of documents and a set of classes, and I must classify these docs in
these classes. Now, documents are stored in Lucene index through Document,
while I don't know how I can store my classes in Lucene, and
Hi,
I'm building an application that's going to classify some documents. So i have
a set of documents and a set of classes, and I must classify these docs in
these classes. Now, documents are stored in Lucene index through Document,
while I don't know how I can store my classes in Lucene, and
[EMAIL PROTECTED] wrote:
Thank you for quick and detailed answer.
In this system multiple threads will, occasionally, try to write and/ or
read the same index, hence the pause waiting for the lock. This is not a
good way to implement it and was done as a temp solution for debug
purposes only.
Here is my source code where I convert pdf files to text for indexing, I
got this source code from lucene in action examples and adapted it for my
convenience, I hop you could help me to fix this problem, anyway if you know
another more efficient way to do it please tell me how to:
import
Aaron,
When you download Lucene from one of the mirrors
http://www.apache.org/dyn/closer.cgi/lucene/java/ (you are using the Java
version, right?), you should see packages named lucene-core-2.0.0.jar. These
contain all Lucene modules and other components that became standard. You need
the
Hi,
Thanks for the reply, I only measure search(), I cached
IndexSearcher in memory.
Best regards, Lisheng
-Original Message-
From: Daniel Naber [mailto:[EMAIL PROTECTED]
Sent: Tuesday, December 05, 2006 12:22 AM
To: java-user@lucene.apache.org
Subject: Re: Lucene search performance:
Hello Lisheng,
a search process has to do usually two thinks. First it has to find the
term in the index. I don’t know the implementation of finding a term in
Lucene. I hope that the index is at least a sorted list or a binary
tree, so it can search binary. The time finding a term depends of
Hello!
I wrote a custom analyzer that has synonyms of some words to help on search.
I use the analyzer when searching the user's entered keyword.
What is happening that I don't understand why is that when tokens are
returned from the synonyms set, the query parser returns the query
: I search my synonyms set and if I find something I return the token like:
: return new Token(synonyms[i], token.startOffset(), token.endOffset(),
: token.type());
: And when it gets do the query I see:
:
: content:wind window
When you add your synonym, it's just going into the stream of
Sorry, I forgot to include this information:
Doing:
token.setPositionIncrement(0);
It returns
content:(wind window)
With:
token.setPositionIncrement(1);
Returns:
content:wind window
I really don't get it..
-Original Message-
From: Chris Hostetter [mailto:[EMAIL PROTECTED]
Sent:
On Tuesday 05 December 2006 21:37, Alice wrote:
It does not work.
Even with the synonyms indexed it is not found.
So if your text contains wind it is not found by the query that prints as
content:(wind window)? Then I suggest you post a small test case that
shows this problem. As Chris
Sounds a very simple and typical use case for a product catalog
search. You are welcome to try DBSight, which is a J2EE web server
that has UI and wizards for you to select data, configure search, and
can run as a production-level search server.
--
Chris Lu
-
Instant
On Tuesday 05 December 2006 20:14, Alice wrote:
It returns
content:(wind window)
That might be the correct representation of a MultiPhraseQuery. So does
your query work anyway? It's just that you cannot use QueryParser again to
parse this output (similar to some other queries like
Hi,
a query like
(-merkel) AND schröder
is parsed as
+(-body:merkel) +body:schröder
I get no hits for this query because +(-body:merkel) doesn't return any
hits (it's not a valid query for Lucene). However, a query like
-merkel AND schröder
works fine. From the user's point-of-view, both
It does not work.
Even with the synonyms indexed it is not found.
That's why my guess was to remove the but I dont know how.
-Original Message-
From: Daniel Naber [mailto:[EMAIL PROTECTED]
Sent: terça-feira, 5 de dezembro de 2006 18:34
To: java-user@lucene.apache.org
Subject: Re:
Just took a quick peak at the MultiPhraseQuery toString() and it does
indeed wrap the query in quotes (it also puts in the parenthesis). You
are generating a MultiPhraseQuery. Is that not your intent?. The
QueryParser will generate a MultiPhraseQuery when more than one token
with different
Ok,
This is the method that adds the aliases, it is located in my SynonymFilter:
private void addAliasesToStack(Token token)
{
String[] synonyms = engine.getSynonyms(contents,
token.termText());
if (synonyms == null)
{
return;
}
Hi Soeren,
Thanks very much for explanations, yes, there
is no linear relation when searching a keyword
which is only in a few docs.
Best regards, Lisheng
-Original Message-
From: Soeren Pekrul [mailto:[EMAIL PROTECTED]
Sent: Tuesday, December 05, 2006 10:37 AM
To:
No.. I am not indexing and searching with the same analyzer.
The reason I do this is because I want to index exactly the contents I have
in my database.
This is used to find some products the company sells, and the users dont
write their names correctly, so if they type something that is
: works fine. From the user's point-of-view, both queries should return the
: same result set. One solution I see is to add a MatchAllDocsQuery clause
: to all prohibited clauses in QueryParser's getBooleanQuery() method. Is
: that a valid solution? I tried with some simple cases and it seems to
As stated before, a *self contained* test case would help people diagnose
your problem ... just cutting and pasting a few snippets of your code is
not enough for people to reproduce your problem.
: And the return is: contents:(wind window)
a MultiPhraseQuery that looks like that should be
I have finally delved back into the Lucene Query parser that I started a
few months back. I am very closing to wrapping up it's initial
development. I am currently looking for anybody willing to help me out
with a little testing and maybe some design consultation (I am not happy
with the
30 matches
Mail list logo