nutch-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrzej Bialecki (JIRA)" <j...@apache.org>
Subject [jira] Commented: (NUTCH-573) Multiple Domains - Query Search
Date Wed, 14 Nov 2007 11:05:43 GMT

    [ https://issues.apache.org/jira/browse/NUTCH-573?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12542414
] 

Andrzej Bialecki  commented on NUTCH-573:
-----------------------------------------

I hope I didn't come across as arguing - your patch looks good from the technical point of
view, I'm just trying to figure out the long-term impact of this patch.

I agree, the full Lucene syntax is too complex - but even the Google syntax falls into the
"advanced" category, i.e. you need to learn how to construct such query. As far as I could
determine, Google indeed treats an infix comma as a list operator but only for some fields,
such as inurl:. Try the following queries:
{code}
site:www.apache.org server
site:www.cnn.com server
site:www.apache.org,www.cnn.com server
{code}

For other fields, such as intitle, inanchor it gives inconsistent results (maybe I discovered
a Google bug :) ).

Regarding the question whether to enable it for any field: I think one important exception
would be "raw fields", where a QueryFilter implementation wants to interpret the input token
differently, and in such cases infix comma may be a valid token character. Perhaps we could
add support for an escape character, which turns comma into a regular token character?

> Multiple Domains - Query Search
> -------------------------------
>
>                 Key: NUTCH-573
>                 URL: https://issues.apache.org/jira/browse/NUTCH-573
>             Project: Nutch
>          Issue Type: Improvement
>          Components: searcher
>    Affects Versions: 0.9.0
>         Environment: All
>            Reporter: Rajasekar Karthik
>            Assignee: Enis Soztutar
>             Fix For: 1.0.0
>
>         Attachments: multiTermQuery_v1.patch
>
>
> Searching multiple domains can be done on Lucene - nut not that efficiently on nutch.
> Query:
> +content:"abc" +(site"www.aaa.com" site:"www.bbb.com")
> works on lucene but the same concept does not work on nutch.
> In Lucene, it works with 
> org.apache.lucene.analysis.KeywordAnalyzer
> org.apache.lucene.analysis.standard.StandardAnalyzer 
> but NOT on
> org.apache.lucene.analysis.SimpleAnalyzer 
> Is Nutch analyzer based on SimpleAnalyzer? In this case, is there a workaround to make
this work? Is there an option to change what analyzer nutch is using? 
> Just FYI, another solution (inefficient I believe) which seems to be working on nutch
> <query> -site:"ccc.com" -site:"ddd.com" 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message