Mailing-List: contact dev-help@jackrabbit.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@jackrabbit.apache.org
Received-SPF: pass (asf.osuosl.org: domain of piccinatto@ibest.com.br
 designates 200.181.68.91 as permitted sender)
Content-Disposition: inline
Content-Transfer-Encoding: binary
Content-Type: text/plain
MIME-Version: 1.0
From: "hsp" <piccinatto@ibest.com.br>
To: dev@jackrabbit.apache.org
Cc: Marcel Reutegger <marcel.reutegger@gmx.net>
Subject: Re: Searching....
Date: Wed, 31 May 2006 12:01:08 -0300
Message-Id: <20060531150108.E79582BB2B@webemail03.ibest.com.br>

>hsp wrote:
>> Normally I put the file inside a node (nt:resource) that is inside a 
>> nt:folder, and do the query:
>> SELECT * FROM nt:folder WHERE CONTAINS(*,'TEXTSEARCHINGFOR')
>
>this will do a fulltext search on nt:folder nodes but not on the resources 
>that are under those nodes, therefore it will never return results.
Ok, but I did for nt:resource too.

>
>> My repository xml does reference to
>> textfilters. The textfilters are based on the information in the
>> jcr:mimetype to build  the index, no?
>
>yes, this is correct.
>
>> Is there no way to rebuild the indexes without have to restart the
>> jboss? I am not seeing this funcionallity works, but I need it working,
>> what are  the best pratice for that (if it exists).
>
>no this is not possible. if you think your index becomes inconsistent 
>during normal operation then it is a bug. please post a jira issue with 
>steps how to reproduce. if you see any errors or warnings in your log 
>files, please also post those. Thanks
I inserted a file.zip tree, and inside it has two files .doc. I can't get the files .doc searching by sql. Otherwise, if I insert just one file .doc per time the sql results the file containg the text.
The operation I do with the zip file is unzip it and insert nt:folder/nt:file like they are in the zip file.

If I delete the tree files of repository to rebuild the indexes, appears the following log:
 11:49:46,803 INFO  [STDOUT] 11:49:46,803 INFO  [RepositoryImpl] Starting repository...
11:49:48,022 INFO  [STDOUT] 11:49:48,022 INFO  [RepositoryImpl] initializing workspace 'default'...
11:49:48,247 INFO  [STDOUT] 11:49:48,247 INFO  [RepositoryImpl] workspace 'default' initialized
11:49:53,502 INFO  [STDOUT] 11:49:53,502 INFO  [SearchIndex] Index initialized: escriba_virtual_dir/repository/index
11:50:05,469 INFO  [STDOUT] 11:50:05,469 INFO  [IndexMerger] merged 1000 documents in 322 ms into _a.
11:50:23,597 INFO  [STDOUT] 11:50:23,597 WARN  [LazyReader] exception initializing reader org.apache.jackrabbit.core.query.MsWordTextFilter$1: java.io.IOException: block[ 0 ] already removed
11:50:46,179 INFO  [STDOUT] 11:50:46,178 INFO  [IndexMerger] merged 1000 documents in 159 ms into _l.
11:51:03,001 INFO  [STDOUT] 11:51:03,001 INFO  [IndexMerger] merged 1000 documents in 112 ms into _w.
11:51:23,566 INFO  [STDOUT] 11:51:23,566 WARN  [LazyReader] exception initializing reader org.apache.jackrabbit.core.query.MsWordTextFilter$1: java.io.IOException: block[ 0 ] already removed
11:51:23,905 INFO  [STDOUT] 11:51:23,904 INFO  [IndexMerger] merged 1000 documents in 92 ms into _17.
11:51:39,166 INFO  [STDOUT] 11:51:39,166 INFO  [IndexMerger] merged 1000 documents in 124 ms into _1i.
11:51:48,953 INFO  [STDOUT] 11:51:48,953 INFO  [IndexMerger] merged 1000 documents in 126 ms into _1t.
11:51:59,477 INFO  [STDOUT] 11:51:59,477 INFO  [IndexMerger] merged 1000 documents in 88 ms into _24.
11:52:10,130 INFO  [STDOUT] 11:52:10,130 INFO  [IndexMerger] merged 1000 documents in 85 ms into _2f.
11:52:30,731 INFO  [STDOUT] 11:52:30,731 INFO  [SearchIndex] Index initialized: escriba_virtual_dir/workspaces/default/index
11:52:30,731 INFO  [STDOUT] 11:52:30,731 INFO  [RepositoryImpl] Repository started


What could be the warning message about LazyReader?
Even after rebuild the indexes, I just find only one .doc file, that is the file I put separated (it is the same that is inside the zip file with the same properties inside a nt:resouce node).

>
>regards
>  marcel
>