You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@nutch.apache.org by kevin chen <ke...@bdsing.com> on 2007/08/26 18:26:03 UTC

search by field

Hi, guys

I am very puzzled by the searching by field.

I can successfully do following query:

site:lucene.apache.org "apache"
which returned 20 hits.

But, following query retured 0 hit.

date:20070101-20070701 "apache"

Using luke, query the same index using query
date:20070101-20070701 apache
returned 2 documents.

What did I miss here? is there some logics implemented for "site:"
keyword that I have to do for other keyword?

--Kevin



Re: search by field

Posted by kevin chen <ke...@bdsing.com>.
Thanks for all the ideas posted back.

Follow Brette_Marc@emc.com's hint, after I enabled plugin query_more,
everything worked well. 

I had only enabled plugin index-more which only cover the index part. To
search by date, query_more has to be enabled also.  I thought any
indexed field should be searchable like luke did, obviously I was wrong.

But, does this mean if I want to add any additional search field that
not covered in index_more, I have to write two plugins, one for index
and one for query?  It seems to me this is a very common task, would be
nice if it's a configurable feature.


On Sun, 2007-08-26 at 21:15 -0400, Erick Erickson wrote:
> Your first query searches over the field site for
> lucene.apache.org and apache.
> 
> Your second query searches over the field date
> for all those terms, and I assume apache isn't 
> in your date field.
> 
> Luke use a default field a bit differently, and
> searched your dates in the date field and
> whatever your default field was for apache. The
> "update" button in Luke is very helpful here, as 
> is the toString() method on queries in your program.
> 
> The toString() method is what I go to first when
> queries don't give me what I expect.......
> 
> Best
> Erick
> 
> 
> On 8/26/07, kevin chen <ke...@bdsing.com> wrote:
>         Hi, guys
>         
>         I am very puzzled by the searching by field.
>         
>         I can successfully do following query:
>         
>         site:lucene.apache.org "apache"
>         which returned 20 hits. 
>         
>         But, following query retured 0 hit.
>         
>         date:20070101-20070701 "apache"
>         
>         Using luke, query the same index using query
>         date:20070101-20070701 apache
>         returned 2 documents.
>         
>         What did I miss here? is there some logics implemented for
>         "site:" 
>         keyword that I have to do for other keyword?
>         
>         --Kevin
>         
>         
> 


Re: search by field

Posted by Erick Erickson <er...@gmail.com>.
Your first query searches over the field site for
lucene.apache.org and apache.

Your second query searches over the field date
for all those terms, and I assume apache isn't
in your date field.

Luke use a default field a bit differently, and
searched your dates in the date field and
whatever your default field was for apache. The
"update" button in Luke is very helpful here, as
is the toString() method on queries in your program.

The toString() method is what I go to first when
queries don't give me what I expect.......

Best
Erick


On 8/26/07, kevin chen <ke...@bdsing.com> wrote:
>
> Hi, guys
>
> I am very puzzled by the searching by field.
>
> I can successfully do following query:
>
> site:lucene.apache.org "apache"
> which returned 20 hits.
>
> But, following query retured 0 hit.
>
> date:20070101-20070701 "apache"
>
> Using luke, query the same index using query
> date:20070101-20070701 apache
> returned 2 documents.
>
> What did I miss here? is there some logics implemented for "site:"
> keyword that I have to do for other keyword?
>
> --Kevin
>
>
>