Hi Gary,
Yes, that certainly helped. Thanks for your reply.
Personally for me, it is interesting to hear about Grails, Ruby etc. I
usually only see folks using PHP, CGI etc. But, maybe it's just me. I
felt like I was being a little anachronistic in using JSP/Servlets.
Thanks again.
Anand
On W
I am having a problem where I am getting lock timeouts when trying to write to
my index file. It would be nice if I could turn on logging to see which
server/application has the lock and when. Is there a way to see the lock
information without changing code?
Thanks,
Billy
___
Robert, I tried to use the worddelimiterfilterfactory as well, but I faced
the same problem[I saw solr using the factory instead of the filter]. I
think I sould try using the first option[copyting that to local directory]
and use it with the options you mentioned. I'll try it out and will post it
h
Hi
One would indeed expect that JAHN would come before JOHNSON. I can't
spot anything wrong with your code but it isn't all there and the
problem could lie with something not shown.
Why don't you cut it down to a nice short self-contained program or
test case and if that doesn't help you find t
kk an easier solution to your first problem is to use
worddelimiterfilterfactory if possible... you can get an instance of
worddelimiter filter from that.
thanks,
robert
On Fri, Jun 5, 2009 at 10:06 AM, Robert Muir wrote:
> kk as for your first issue, that WordDelimiterFilter is package
> protect
kk as for your first issue, that WordDelimiterFilter is package
protected, one option is to make a copy of the code and change the
class declaration to public.
the other option is to put your entire analyzer in
'org.apache.solr.analysis' package so that you can access it...
for the 2nd issue, yes
Can somebody take a look in why I'm not getting the correct alphabetical
order?
I changed last_name field to be 'UN_TOKENIZED' and sorting on only last
name, then network status. But with last name as 'J' in query, I am getting
results like:
JACOBSON
JOHNSON
JAHN
I would expect 'JAHN' to come b
Enumerating terms will be inefficient compared to getting the stored field.I'd
try soring the fields first until and unless you can demonstrate a problem.
BTW, if you're not going to *search* on the field, there's no reason to
index
it at all.
Why do think you don't want to store the paths? How bi
Thanks Robert. There is one problem though, I'm able to plugin the word
delimiter filter from solr-nightly jar file. When I tried to do something
like,
TokenStream ts = new WhitespaceTokenizer(reader);
ts = new WordDelimiterFilter(ts);
ts = new PorterStemmerFilter(ts);
...rest as in the l
KK, you got the right idea.
though I think you might want to change the order, move the stopfilter
before the porter stem filter... otherwise it might not work correctly.
On Fri, Jun 5, 2009 at 8:05 AM, KK wrote:
> Thanks Robert. This is exactly what I did and its working but delimiter is
> mi
Thanks Robert. This is exactly what I did and its working but delimiter is
missing I'm going to add that from solr-nightly.jar
/**
* Analyzer for Indian language.
*/
public class IndicAnalyzer extends Analyzer {
public TokenStream tokenStream(String fieldName, Reader reader) {
TokenStream
i think you are on the right track... once you build your analyzer, put it
in your classpath and play around with it in luke and see if it does what
you want.
On Fri, Jun 5, 2009 at 3:19 AM, KK wrote:
> Hi Robert,
> This is what I copied from ThaiAnalyzer @ lucene contrib
>
> public class ThaiAn
Say I have indexed the following strings:
1. "cool gaming laptop"
2. "cool gaming lappy"
3. "gaming laptop cool"
Now when I search with a query say "cool gaming computer", I want string 1
and 2 to appear on top (where search terms are closer to each other)
followed by 3.
I can use a Term query t
You can't get at the field values from the document hits.
Field.Store.NO means it isn't stored and what isn't there can't be
retrieved.
But you should be able to get at the indexed paths via a TermEnum.
Something like this
IndexReader reader = IndexReader.open(...);
String field =
Hi
My guess is that one big index would be more efficient since the total
IO read and write load would be less. The big reason for creating
smaller intermediate indexes is that you could spread their creation
over multiple jobs/disks/servers.
There is lots of good advice in
http://wiki.apache.o
Hello,
I am wondering whether using Lucene Highlighter it is possible to highlight
parts of words. So for instance, if I type A then in word America I would
have America or if I type er I would get America
Thanks for replies
--
View this message in context:
http://www.nabble.com/One-character-
Thanks for the up Otis. I will give this some more thought, prototype
some, and possibly put in a proposal for the Apache Incubator.
Ye,
I am not aware of Sixearch, but there are several P2P applications
e.g. WiredReach, Grub, Neurogrid etc.
However, my idea is quite a bit different from the exis
Hi Robert,
This is what I copied from ThaiAnalyzer @ lucene contrib
public class ThaiAnalyzer extends Analyzer {
public TokenStream tokenStream(String fieldName, Reader reader) {
TokenStream ts = new StandardTokenizer(reader);
ts = new StandardFilter(ts);
ts = new ThaiWordFilter(ts
18 matches
Mail list logo
