lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Erik Hatcher <e...@ehatchersolutions.com>
Subject Re: Range Query Sombody HELP please
Date Mon, 31 May 2004 13:20:28 GMT
Try my AnalysisDemo code on some filename field samples:

	http://wiki.apache.org/jakarta-lucene/AnalysisParalysis

You mentioned earlier, I think, that you are using a custom analyzer.  
Give us the output of AnalysisDemo on some samples so we can see what 
is coming out.

If you can put together a 10-line Java program that uses RAMDirectory 
and has some sample hard-coded text that I can easily run standalone I 
would look into your situation further.  As it is, you are providing 
far more complexity than I have time to delve into.  Narrow it down to 
a very very simple example that we can all see in one screen.

	Erik


On May 31, 2004, at 7:47 AM, Karthik N S wrote:

> Hey Ype...
>
> 1) I switched Off the Multi search Senerio.
>
> 2) Changing the Field type from Text to Keyword
>     will fail When I search for the the Field type  "filename"
>     so,I still maintained it to be Text
>
> D:\JAVA\lucene\src\demo>java org.lucene.src.indexer.search.SearchFiles
> Search Keyword : b10181_p388
> Source path [ E:/po/aaaa ] : e:/indexer3/b10181
> Query: ['b10181_p388'] in Folder e:/indexer3/b10181/b10181_indx_
>
> Found document(s) that matched : 'b10181_p388' no of hits :'1' in query
> Field :'filename'
> File Name     : B10181_P388
>
>
> 3)On Search for range between 2 file names  B10181_P702   to  
> B01081_P355
>     still returns me  0 hits  [Included space before the 2nd '+' ]
>
> D:\JAVA\lucene\src\demo>java org.lucene.src.indexer.search.SearchFiles
> Search Keyword : +button +filename:[b10181_p702 TO b10181_p355]
> Source path [ E:/po/aaaa ] : e:/indexer3/b10181
> Query: ['+button +filename:[b10181_p702 TO b10181_p355]'] in Folder
> e:/indexer3/b10181/b10181_indx_
> Not a Found document(s) that matched query Field 'filename':
>
> 				or
>
> D:\JAVA\lucene\src\demo>java com.controlnet.indexing.search.SearchFiles
> Search Keyword : +contents:button +filename:[b10181_p702 TO 
> b10181_p355]
> Source path [ E:/po/aaaa ] : e:/indexer3/b10181
> Query: ['+contents:button +filename:[b10181_p702 TO b10181_p355]'] in 
> Folder
> e:/indexer3/b10181/b10181_indx_
> Not a Found document(s) that matched query Field 'filename':
>
>
> Also the does the search varies on the Field Type if so My Indexed 
> Field
> types as below....
>
> doc.add(Field.Text("path", fhtml.getPath()));
> doc.add(Field.Keyword("modified",fhtml.lastModified()+""));
> doc.add(Field.Text("filename",fhtml.getName()));
> doc.add(Field.Keyword("creation",CREATION_));
> doc.add(Field.Keyword("bookid",BOOKID_));
> doc.add(Field.Text("chapNme",CHAPNAME_));
> doc.add(Field.Text("itmName",ITEMNAME_));
>
>
>
> please do advise me.
> Karthik
>
>
>
> [ James Goslink says   Microsoft has More Money to burn then GOD has
>   ...on his visit to India,In an interview to MSNBC TV Last night ]
>
>
>
> -----Original Message-----
> From: Ype Kingma [mailto:ykingma@xs4all.nl]
> Sent: Monday, May 31, 2004 2:52 PM
> To: lucene-user@jakarta.apache.org
> Subject: Re: Range Query Sombody HELP please
>
>
> On Monday 31 May 2004 11:09, Karthik N S wrote:
>
> ...
>> I re indexed my folder 10181 [Seem's to be corrupted]
>
> Was the index writer closed?
>
>> Now I am getting the hits as....
>>
>>
>> D:\JAVA\lucene\src\demo>java org.lucene.src.indexer.search.SearchFiles
>> Search Keyword : +button+filename:[B10181_P702 TO B01081_P355]
>
> The query needs to have space before the 2nd + :
>
> +button  +filename:[B10181_P702 TO B01081_P355]
>
>> Source path [ E:/po/aaaa ] : e:/indexer3/b10181
>> Query: ['+button+filename:[B10181_P702 TO B01081_P355]'] in Folder
>> e:/indexer3/b10181/b10181_indx_
>> Not a Found document(s) that matched query Field 'filename':
>> Not a Found document(s) that matched query Field 'bookid':
>> Not a Found document(s) that matched query Field 'creation':
>> Not a Found document(s) that matched query Field 'contents':
>> Not a Found document(s) that matched query Field 'chapNme':
>> Not a Found document(s) that matched query Field 'itmName':
>
> You seem to use a search mechanism that searches all these fields.
> I'd recommend to switch this off until a query with explicit fields 
> works,
> eg.:
>
> +contents:button  +filename:[B10181_P702 TO B01081_P355]
>
> Btw. You'll need to make sure that a term like B10181_P702 is
> not split at the underscore _ by a tokenizer at indexing time.
> If your filename is not a keyword field, you might consider
> changing it into a keyword field.
>
> You seem to index book pages as Lucene documents, which is ok.
> However, you may also need to index larger parts of the books in
> order to retrieve books with multiple subjects on different pages.
> Is this what your original question is about?
>
> Have fun,
> Ype
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: lucene-user-help@jakarta.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: lucene-user-help@jakarta.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: lucene-user-help@jakarta.apache.org


Mime
View raw message