Hi!
Imagine an index holding documents in different languages and country.
Language+country is what I call a context and I build and hold a QueryFilter
for each context.
When performing a fuzzy search, FilteredTermEnum doesn't care about any
contexts at all (well, how should it :). It builds a
Hi!
I search an 1.5 gig index and fuzzy queries are really slow; something like
avg. ~500ms (IndexSearcher.search(Query, HitCollector)).
When performing exact queries I archieve response times <25ms. What is it that
makes fuzzy queries so slow? Increased index access due to more terms, i.e.
d
On Saturday 24 November 2007 18:28:48 markharw00d wrote:
> The added IO is one factor. Another is the CPU load from doing many
> edit-distance comparisons between index terms and the provided search
You mean FuzzyQuery.rewrite(). Are you sure this is a CPU and not an IO issue
(reading the terms f
might be in a document or field that I'm not interested in at
all.
> you first select used word wich share ngram with the query word, the
> distance is computed with levenstein, and you use this word as a
> synonym.
>
> M.
>
> Le 24 nov. 07 à 17:36, Timo Nentwig a écrit :
>
On Saturday 24 November 2007 18:28:48 markharw00d wrote:
> term. You can limit the number of edit distance comparisons conducted by
> setting the minimum prefix length. This is a property of the QueryParser
Well, javadoc: "prefixLength - length of common (non-fuzzy) prefix". So, this
is some kind
d then if you're lucky files==philes but there's no
> room for error and they either match or they dont - there is no measure
> of similarity.
>
> There's no free lunch here.
>
> Timo Nentwig wrote:
> > On Saturday 24 November 2007 18:28:48 markharw00d wrote:
On Friday 30 November 2007 12:59:13 Grant Ingersoll wrote:
> Hmmm, I think you should be able to rely on the fact that Fields are
> stored in order of indexing and then read back in that same order.
Yeah, tought about that for a moment but this is just way to fragile.
> Otherwise, the reading twi
Hi!
I do have different document types (Books, Magazines, Author whatever) in the
index and a FieldSelector is document type specific (for Books LOAD isbn and
title for Author name, ...). The document type can be determined by a field
surprisingly called documentType.
How am I going to do this
On Friday 30 November 2007 19:28:12 Grant Ingersoll wrote:
> I guess the question becomes what is the nature of your fields? Do
> you have some really large fields that you want to avoid loading b/c
> they are not shown initially? That is the main use case, I guess.
I wonder why there's not Lazy
On Wednesday 05 December 2007 12:20:51 Grant Ingersoll wrote:
> Then, when you go to access those 4 fields, which you most certainly
> will at some point soon, otherwise why did you get the document to
Nope, I won't :)
In fact my Document contain fields I only need for searching and sorting. But
Hi!
Is there are particular reason why CachingWrapperFilter caches per IndexReader
and not per IndexReader.directory()? If there are multiple
IndexSearcher/IndexReader instances (and only one Directory) cache will be
built and held in memory redundantly. I don't see any sense in doing so (?).
PROTECTED]> wrote:
> > My guess would be b/c best practice is usually to only have one Reader/
> > Searcher per Directory, but I don't know if that is the real reason.
> > Most discussions/testing I have seen indicate a single Reader/Searcher
> > performs best.
> &
since does a single thread plus synchronous IO scale?
However I yet didn't find any discussion on this topic so I'd be glad if
somebody could give me a link.
> -Grant
>
> On Jan 1, 2008, at 11:57 AM, Timo Nentwig wrote:
> > Hi!
> >
> > Is there are particular reaso
On Tuesday 01 January 2008 21:06:06 Mark Miller wrote:
> The main reason to use a single IndexReader is because its very time
> consuming to open an IndexReader. If your index is pretty static, maybe
Yes, it takes quite some time to build it and it's not changed but rebuilt
from scratch.
> Perha
fferent HDs but *of couse* we're talking
about multiple hard drives (at least some RAID, in my case it's some
expensive netapp, however I don't know which one exactely but I can find
out...).
> even better, separate systems (using RMI or something).
>
> - Mark
>
> Timo
On Wednesday 02 January 2008 08:03:48 Chris Hostetter wrote:
> 1) there is a semi-articulated goal of moving away from "under the
> coveres" weakref caching to more explicit and controllable caching ...
YES!
BTW why havin caching been removed from QueryFilter at all? Isn't caching the
only sens
Hi!
I found that when sorting the search result -depending on the amount of data
in the field to sort by - this can easily lead to FieldCacheImpl to allocate
hundreds of MByte RAM.
How does this work internally? It seems as if all data for this field found in
the entire index is read into memo
ric field, make
that explicit.
Otis
--
Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch
- Original Message ----
From: Timo Nentwig <[EMAIL PROTECTED]>
To: java-user@lucene.apache.org
Sent: Sunday, April 13, 2008 4:45:37 PM
Subject: Sorting consumes hundreds of MBytes RAM
Hi!
I
Hi!
I advised Directory.close() (AspectJ) and noticed that it's not called at all
for the following code:
final FSDirectory d = FSDirectory.getDirectory( path );
final IndexReader r = IndexReader.open( d );
final IndexSearcher s = new IndexSearcher( r );
...
s.close();
r.close();
d.close(); // D
Hi!
Unfortunately the search method in ParallelMultiSearcher which is able to take
an HitCollector isn't running in parallel and there' even an issue regarding
this (LUCENE-990) with zero watchers or votes :-\
So this isn't something that's likely to be done in near future, is it?
And question
Hi!
Since Filter.bits() is deprecated and replaced by getDocIdSet() now I wonder
how I am supposed to combine (AND) filters (for facets).
I worked around this issue by extending Filter and let getDocIdSet() return an
OpenBitSet to ensure that this implementation is used everywhere and casting
in two classes that do (precisely?) what you need:
> contrib/miscellaneous/**/ChainedFilter
> contrib/queries/**/BooleanFilter
>
> Regards,
> Paul Elschot
>
> Op Saturday 08 November 2008 19:06:15 schreef Timo Nentwig:
> > Hi!
> >
> > Since Filter.bits() is dep
Hi everybody,
is it possible to store fields without term position (the .prx file) data? We
store sort of custom
data in the field and use it as some sort of a filter for queries, so we just
don't need any term
position data and it bloats the index' size nearly by factor 3.
Thanks
Timo
---
23 matches
Mail list logo