In the sample TestSort represents my problem:
In the info below I need get the list of "contents" that contains
"x" (A,C,E,G,I) and other list of index (5,2,3) that not contain info
replicated.
The first list I get using any query of type: query = new TermQuery
(new Term ("contents", "x"))
German,
How would be it ?
You have 2 index ?. One for seach main (keyword) and other for location ?
You do 2 search, The first is the search main e the second is the
search location ,but insert the filter.
What type of Filter do use ?
I have the bitset of search main (keyword), but I
A facet is a group condition, could be a single value of the doc or a
set of filters.
On Nov 19, 2007 1:09 PM, Haroldo Nascimento <[EMAIL PROTECTED]> wrote:
> German,
>
> When You said:
> "I collect every facet's bitset ... "
> what is a facet ? Is there the each option of filter of your site ?
I have already defined a Lucene Filter for every "id" of "ubicacion".
I just create the bitset for every value, and count it against the result.
One possible optimization is to read the terms of the field you're
trying to "group", that's the optimization we'll be working soon on
our app.
I never
German,
When You said:
"I collect every facet's bitset ... "
what is a facet ? Is there the each option of filter of your site ?
How you get the every facets ?
On Nov 19, 2007 1:05 PM, Haroldo Nascimento <[EMAIL PROTECTED]> wrote:
> German,
>
> What I need is similar to the your site
> ht
German,
What I need is similar to the your site
http://listados.deremate.com.ar/panaderia .
I have many results of search, but I show any result (for example:
first 10 for first page) , but for create the options of filter of
location I need read all results fof search. The problem of
performa
I think, based on your previous question, that you just need to use
the search() method that returns TopDocs, not the lower-level
HitCollector method. From the TopDocs, you can then access the
ScoreDoc, which will give you info about the doc and the score. See http://www.lucenebootcamp.com/
Why do you need the doc's info?
If you're grouping you may not need detail on each group condition.
Here is a sample of faceted (grouped) search:
http://listados.deremate.com.ar/mp3
(Sorry, it's in spanish)
Simply I collect every facet's bitset and intersect it against the
result's bitset (keywo
Mark,
How I can get the information of Document. I think that is in the
implementation do method abstract collect. How I can get it .
Below is the example of javadoc the Lucene.
Searcher searcher = new IndexSearcher(indexReader);
final BitSet bits = new BitSet(indexReader.maxDoc());
se
You sould never use the hits for other use than retrieving a group of
results (usually a page of 10-20-30 docs).
You could see Apache Solr's implementation of faceted search.
I've use that code as a guide to group & count diferent facets (or
conditions, fields as you wanna call it), is pretty fast
You can feed the hits vector into Quaere (http://quaere.codehaus.org/) to
accomplish the SQL-like grouping you desire, very easily. But I'm not sure
it'll be that much quicker. Worth a shot.
T
On 11/18/07, Haroldo Nascimento <[EMAIL PROTECTED]> wrote:
>
> I have a problem of performance when I ne
Hey!
Search for the topic "Aggregating Category Hits" in the list. You'll get
a few approaches that you may use to implement "groupby".
Regards,
kapilChhabra
-Original Message-
From: Haroldo Nascimento [mailto:[EMAIL PROTECTED]
Sent: Monday, November 19, 2007 3:02 AM
To: java-user@lucene
Correction: that issue to watch out for is in regards to the TopDocs
HitCollector. If you where to go with your own HitCollector rather than
TopDocs you might not necessarily have this problem (or at the least you
can code around it).
Mark Miller wrote:
Hey Haroldo.
First thing you need to d
Hey Haroldo.
First thing you need to do is *stop* using Hits in your searches. Hits
is optimized for some pretty specific use cases and you will get along
much better by using a HitCollector.
Hits has three main functions:
It caches documents, normalizes scores, and stores ids associated wit
Can you explain the problem you're trying to address from the user's
perspective?
From the description you've provided, you may want to look up
"Faceted Searching". Another option may be to use a HitCollector,
but it would help us if you could describe the problem at a higher
level.
Re
15 matches
Mail list logo