: 1-15. On first inspection I tried a RangeQuery but, of course,
: I quickly discovered that RangeQueries expand to boolean queries that
: enumerate every element in their range. In my case, it does not seem
: realistic to enumerate all the possible values of the query range as
: RangeQue
Any term will do, although I'd personally choose a very common one.
Otis
--- Cheolgoo Kang <[EMAIL PROTECTED]> wrote:
> Thanks Otis,
>
> I'm doing exactly what you've said, except the sorting. I'll try it
> out immediately.
>
> Anyway, will the warming up search with ANY term do the same job?
Thanks Otis,
I'm doing exactly what you've said, except the sorting. I'll try it
out immediately.
Anyway, will the warming up search with ANY term do the same job? Or
should I have to search with the worst term that has the longest hits?
On 12/11/05, Otis Gospodnetic <[EMAIL PROTECTED]> wrote:
>
Hello and Good Day,
In my application of Lucene, I am must search through some fields that
contain numbers with very large ranges on the order of 15 or so.
Suppose I wanted to retrieve all documents matching numbers
1-15. On first inspection I tried a RangeQuery but, of course,
Yes you can. Just open an IndexReader/Searcher, say in a separate
thread, run a search on it, and then swap your previous
IndexReader/Searcher with the warm one. If you use sorting, make your
"background query" with the sort option, so the FieldCaches get
populated and warmed up.
Otis
--- Cheol
Hi,
I'm running an index on FSDirectory with 0.4M documents with each of 7 fields.
When I open an IndexReader and an IndexSearcher, the average search
time with hits of 0.2M items (yeah, very common word) takes about
150~250 msec and it's pretty good. But the first time just after
opening IndexRe
On Dec 10, 2005, at 5:17 AM, Lawrence wrote:
Is there a projected date for the next release?
Real Soon Now :)
I am using the code from SVN repos and I am hoping the next release
will be before I go to production. I always like a little stability
in my life.
You can pretty safely rely on
first thanks for your response Dan,
It I'll be around 5.000 database records with three indexed fields:
id, title(1 line) and description(around three lines). I was even
considering using the in memory feature for faster access but I'm new to
lucene and I'don't know if that I'll cause my proble
If this is a small index and it won't change after install (you are just
using it to search, not to index), place it in a sub-directory of WEB-INF.
If it is a larger index (something you don't want to copy frequently), or it
will change after install, then you shouldn't keep it inside your web
app
What would be the best practice storing the index in a webapp.
I mean in wich folder?
Thanks.
Raul.
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Thanks Chris I appreciate your comments.
I have succesfully fixed the first changing the parameter to true and
the second error implementing three interfaces in one class and adding
some lines to hibernate.cfg.xml, below is the code in case somebody is
interested:
hibernate.cfg.xml***
Hello,
Is there a projected date for the next release?
I am using the code from SVN repos and I am hoping the next release will be
before I go to production. I always like a little stability in my life.
Thanks for all your good work,
Lawrence
-
On 12/10/05, Chris Hostetter <[EMAIL PROTECTED]> wrote:
> : are there any APIs which will index mysql databases and run periodically ?
>
> I'll defer that question to someone else who knows more about it.
>
I may not know more than most people. But to index databases and run
periodically, you need
: are there any APIs which will index mysql databases and run periodically ?
I'll defer that question to someone else who knows more about it.
: i have one more query: if i choose to search on multiple fields do i loose
: the advantage of fuzzy search and stuff like that
Absolutely not. Searchi
: following the interceptor method (the second one in
: http://www.hibernate.org/138.html)
I don't know much about hibernate, but I think i understand both of your
questions...
: 1. Lucene's doesn't create the needed folder structure the first time
: and it complains that it can't find segments.
15 matches
Mail list logo